AI3021- IT in Agriculture

UNIT III
Agricultural systems - managerial overview, Reliability of agricultural systems, Simulation of
crop growth and field operations, Optimizing the use of resources, Linear programming,
Project scheduling, Artificial intelligence and decision support systems.

AGRICULTURAL SYSTEMS MANAGEMENT

AGRICULTURAL SYSTEMS DEFINED

• An agricultural system is a specified group of components, operational functions,

and processes that are integrated to accomplish a well-defined purpose.
• Agricultural systems managers (ASMs) usually plan, evaluate, and adjust a
system or some group of components of a system. In a complex agricultural
system one can easily identify or envision systems within systems.
• We might call those systems within systems subsystems, but the main concern
is that the system performs to the highest level we can achieve within the
parameters or resources at hand.
• The ASM must learn quickly to manage and evaluate subsystems concurrently.
CHARACTERISTICS OF AGRICULTURAL SYSTEMS
• Each agricultural system has some well- defined purpose. The purpose might be
stated simply, or more often it might be stated as a group of objectives.
• And those objectives might include some very definite specifications or
measures of performance.
• As an example, let’s consider an agricultural system that produces potato chips.
The overall purpose might be to produce six 1 2= -ounce bags of chips for the
consumer market.
• However, the well-defined purpose for the line technical mangers of the
processing plant might be production objectives of 1500 bags per hour, less than
0.05% of bags over 0.675 ounces, with less than 8.00% waste, less than 1% under
0.6355 ounces, with less than 0.5 hours line downtime per shift, less than 1
recorded human accident per 6 months, at a cost of less than $0.233 per bag, and
meeting the 2002 EPA guidelines regarding odor control in the community.
• In real life this specification list could be much longer, involving hours of
planning and discussion with managers, engineers, accountants, labour unions,
and lawyers.
We could easily identify the same level of complexity in a production farming
operation. The overall purpose might simply be to grow hogs for meat processing. The
well-defined purpose might be to produce 1000 barrows, with weight range limited to
205–215 lb, loin eyes of not less than 4.14 sq in or
more than 5.01 sq in, average rates of gain of not less than 1.98 lb/day, at a cost of less
than $0.21 per pound, utilizing non GMO feeds, and meeting all local and federal EPA
requirements. And, once again, the real pork producers reading this chapter know that
an actual specification list for this agricultural system would be much longer. This
AI3021- IT in Agriculture
discussion of the ‘‘well-defined purpose’’ portion of the definition is well served by the
two examples. Looking back at our potato chip processing plant example, we know that
to fulfil the well-defined purpose (specifications), numerous operational functions and
processes need to be performed. The operational functions and processes are key parts
of our definition. Some of those operational functions and processes involving our
complete definition of the potato chip agricultural system might be:
• Purchasing of potatoes
• Transportation
• Unloading at the plant
• Storage
• Grading and inspection
• Sorting
• Chemical wash
• Storage
• Waste disposal
• Purchasing of frying oils
• Transportation of oils
• Storage of oils
• Least-cost formulation of ingredients
• Cutting processes
• Line movement of product
• Frying
• Salting
• Packaging/weighing
• Boxing
• Quality control
• Pricing
• Loading
• Distribution and storage
• Sales
• Transport
In this agricultural system each operation or process might require humans, machine
components, or a chemical or biological process. And each process will have
management-defined parameters for successful operation. We could construct elaborate
diagrams or computer programs to inspect or evaluate each component or operational
activity. But our mission is to manage agricultural systems to achieve goals. To do that,
we must fully understand how each component of a system works and the interacting
effect its application might have on other components and the outcome of the system.
To predict those impacts, we can rely on mathematical models and tools to forecast
outcomes of decision alternatives.
AI3021- IT in Agriculture
EXPECTED OUTCOMES OR ‘‘DELIVERABLES’’ OF AGRICULTURAL
SYSTEMS
There is a real need for the system to deliver well-defined products. Today’s
consumer-oriented market demands truth and performance. A system often must meet
standards either the consumer wants or governments might demand. Let us consider the
challenge a small organic fertilizer bagging plant might have. The manager must meet
the desires of the company for profit, of the public for performance, and of the
government for truth in packaging. The manager and the company must design a
product that meets the specification that satisfy company upper management for profit,
establish product features that entice a certain market share of buyers and meets state
and federal analysis on the package label. Each fertilizer product becomes very well
defined, and the agricultural processing system must meet the goals. Many people from
many arenas give critical input.
It is expected that the products the production line delivers are profitable.
Agriculture is a business Every business must in the end meet a profit goal. In this small
organic fertilizer bagging business we must set production volume targets, estimate
fixed and variable costs, and establish costs per bag, pricing, and target profits. If this
was all there was to management of this line, it would be easy. However, agricultural
systems have to satisfy many other conditions. Profitability is not enough. The product
must be safe for use. The manufacturer must also guarantee the safety of the production
workers and perhaps meet union work conditions. If the product were edible or a drug
it would have to be safe for human consumption Certain aspects of quality control
would have to be met and verified.
Our bagging plant would probably produce ‘‘secondary’’ compounds or wastes.
They must be managed as well. Those processes and costs become part of the total
agricultural system too. The plant is responsible in American society to maintain or
enhance the environment. Or at the very least, the product must be produced at
environmental costs that society (and the law) deems acceptable. It is not unusual to
expect that the bagging plant system also contribute to the general community economic
development and well-being.

THE PRIMARY GOALS OF AGRICULTURAL SYSTEMS MANAGERS

An all-too-common mistake of technical management is a failure to identify
performance criteria against which the system must be evaluated. These criteria and
their measures need to be well established before we ‘‘flip the switch’’ and a system
begins to operate.
The goals, criteria for success, and performance measurement assessment should
be part and parcel of every planning process and evaluation process. These performance
criteria should be established up front and agreed upon by all management.
It is not unusual to see these measures as part of monthly, quarterly, and annual
reports. Some primary goals agricultural systems managers must attain are:
1. Optimization of economic costs, profits, and benefits
AI3021- IT in Agriculture
2. Production of defined levels of product quality and
3. quantity
4. Meeting timelines and schedules
5. Delivery of value-added products and product attributes
6. Attaining acceptable process reliability
7. Maximization of e efficiencies
8. Realization of environmental and regulatory guidelines
9. Optimization of human factors—safety, job satisfaction,
10. performance factors, and perhaps labor union issues
CRITERIA OF EVALUATION
Obviously, production costs per bag of chips would be a key criterion to evaluate,
or maybe the total volume over a week, or total production in bags over the year.
So, costs, volumes, and profits are key criteria. Efficiency of the processes can
be measured, equipment adjustments made, or replacement of processes effected.
Some of the processes included truck unloading, storage, belt movement, cutting,
frying, salting, sorting, weighing, bagging, and boxing. Product quality must also be
assessed.
Yes, we all like a beautiful, unbroken chip! The size, colour, weight, etc can all
be assessed. Storage life and condition of the package itself can be evaluated and goals
set No one likes a wrinkled bag or a misprinted label.
And who wants a six-ounce bag of chips labelled as seven ounces? If we forget
that the whole system must have a very high-performance reliability, we cannot achieve
any other goals.
To achieve this, we must meet timelines and schedules and attain serviceability
and machine replacement goals. In a production plant we are always scrutinizing
resource utilization, waste reduction, risk minimization, performance, stability,
environmental impacts, FDA guidelines, food safety guidelines, and state and federal
regulations.
Today, food processes must meet purity and process standards. Then there is the
human element.
Are the plant meeting union agreements? Some additional criteria for evaluation
would be meeting goals in plant safety, personnel satisfaction, health, comfort, and plant
security.
The ASM must approach the system management from a team perspective.
Many have input. Many must be informed and empowered. It becomes quite clear that
a good technical manager manages far more than money.
Managing means managing all the resources. The good technical manager must
also consider machine replacement, new products, new technologies, and training of
personnel.

FACTORS AFFECTING THE AGRICULTURAL SYSTEM AND THE ASM’S

DECISION MAKING

1. Changes in Weather, Seasonality, or Biological Intrusions

Perhaps the most significant and unique factor in agricultural systems
management is dealing with the weather or seasonality of commodity production.
AI3021- IT in Agriculture
In our organic bagging plant example, an unknown or unplanned warm weather
span could greatly spur biological activity of stored wood chips or manure.
Rising temperatures or microbe levels in the potato storage sheds would cause
the potato chip manger to have to adapt to these conditions. Many processed products
of agricultural systems are live biological entities requiring heating, cooling,
pasteurization, sterilization, fermentation, or even radiation.
Seasonal changes or unusual weather patterns can greatly change schedules in
Beld planting of corn or harvesting of soybeans. A meat processor knows that biological
processes occur in known time frames. He or she must acknowledge and respond to
unusual temperatures.
Processing changes must occur, or product quality is lost—or the product itself
could be lost entirely.
A Florida citrus grower must change processes if an unusual cold snap
jeopardizes the life of a young fruit tree Or if a disease such as aflatoxin enters a corn
Beld nearing harvest, an immediate response by the ASM is required.
In production agriculture one must always be prepared to alter decisions when
plant or animal diseases enter or threaten. The threat of SDS (sudden death syndrome)
to a soybean crop would require changes in variety selection, planting dates, and harvest
dates.
Likewise, a dry year would certainly spur the ASM of a grape vineyard to engage
irrigation scheduling. The same dry year would spur the ASM of a winery to change
the formulation of his winemaking process, since the soluble solids count of the grape
juice would increase in a dry year.
In a wet year, the count may drop and the ASM might add sugar to fermentation
processes.

2. CHANGES IN TECHNOLOGY
All so often dramatic changes in technology or innovations impact the agricultural
system so much that the ASM is required to completely change the components,
functions, or processes of the systems. In other words, we completely change the way
we do things.
One dramatic example of this is the impact of biotechnology techniques and
nanotechnology equipment on the development of plant seeds. The new technology
completely changed how we exchange genetic information to form new varieties.
Processes were changed, new skills were required, and old seed technology was
rendered noncompetitive. While this new technology changed forever how we
promulgate plants, it also changed how we grow them in the Beld Genetically modified
plants are now collegial in being resistant to certain herbicides.
Thus, we also have modified the cultivation and pesticide application in the
production Beld systems of agriculture. Changing a technological process is not the
only impact of changing technology. The development of a new product can
greatly change an agricultural system.
The invention of the large round hay baler is a good example. The introduction of the
machine completely revolutionized haymaking in the Midwest, where labour costs are
high. The old system of baling hay in small rectangular bales was rendered
economically noncompetitive, except in specialty markets. Likewise, another
biotechnological breakthrough is allowing us to grow pharmaceutical proteins and
compounds in corn.
AI3021- IT in Agriculture
Several billion dollars a year is now generated by growing this new ‘‘Pharma’’ corn
product, but the system of growing and handling requires new and unusual techniques
to ensure biological security of plant growing regions ‘‘Pharming’’ requires many
changes in the agricultural system.
Sometimes the breakthroughs can come from other industries or other countries.
Agricultural industries and systems were greatly affected when other manufacturing
industries began to adopt and develop different sweeteners.
Corn growers benefited from high-fructose sweetener, while sugar cane growers
were forced to change production methods to remain competitive.
Better irrigation technology in Israel and Brazil forced growers of citrus in
Florida, Arizona, and Texas to completely change irrigation technology to remain
competitive. Some examples of technology changes of great impact
would be:
• Analog/digital interfacing with microcomputers
• Global positioning systems (GPS)
• Introduction of microcomputers for data handling and
• controls
• Spreadsheet software
• New plastic extrusion methods
• Ethanol processing from corn
• Rotary threshing mechanisms in combines
• Ergonomic engineering of tractor cabs
• Soil conservation practices
• The cotton gin
• Evaporative cooling for greenhouses
• Hydra cooling of fruits and vegetables
• Irradiation of meats, fruits, and vegetables
The list is very long and continues to grow daily. One of new technologies having the
most impact is the use of the Internet for marketing and purchasing—commonly called
e-commerce. E-commerce now allows an ASM to purchase and market worldwide. Top
ASMs will need information systems that allow them to be educated rapidly regarding
new developments. The Internet itself is a technological addition that has had perhaps
the most dramatic effect on 21st century agriculture.

3. LEGAL/POLITICAL FACTORS
New laws and regulations can have great impacts on decisions regarding Beld
production, manufacturing and processing, and technical marketing areas of agricultural
systems. Even without new laws, new rulings by regulatory agencies can have
consequences.
Changes in the tax structure can have significant impacts on management.
Throughout the 1970s and early 1980s, farmers enjoyed federal tax exclusions from an
investment credit deduction. Farmers could derive great benefits that encouraged
buying capital equipment such as tractors, combines, and portable buildings. Federal
tax reform removed
these advantages, and equipment replacement planning strategies changed greatly.
Because many did not know their income status until late in the year, there was a lot of
lastminute December purchasing. This last-minute buying ceased. This change affected
AI3021- IT in Agriculture
not only farmer purchasing but the way money was spent. Manufacturing schedules,
technical sales programs and activities, and managerial decisions were changed. Even
tax accountants had to change their schedules of activity.
Some of the most significant laws now affect the livestock production industry.
Some small rural cities now have ‘‘influence’’ up to three miles from their city limits
regarding odor control. Many local agencies in counties now control animal unit limits.
Changes in fees for grazing on public lands in the West are another example where
ranchers are forced to manage differently under different rate structures.
Laws affecting migrant labor, labor camp conditions, and wage rates greatly
affect the fruit and vegetable industries. The trade-off between labor and mechanization
greatly changes.
Likewise, Occupational Safety and Health Administration (OSHA) regulations
and labor laws impact management decisions in processing plants.
On the technical marketing scene, the North American Free Trade Agreement
(NAFTA) has changed the playing Beld considerably. Some industries have greatly
benefited, while others have suffered. Changes in EPA guidelines and standards now
have great managerial impact regarding the use of fertilizers and pesticides.

Nonpoint and point sources of watershed runoff are now more controlled. The ASM
must keep abreast of key local, national, and international issues. Some key
governmental agencies are:

1. Bureau of Land Management

2. EPA
3. State Departments of Agriculture
4. Water management districts
5. Farm Services
6. US Forest Service
7. Zoning commissions
8. Department of the Interior
9. Bureau of Indian AAairs
10. Agricultural Plant Health Inspection Service (APHIS)
11. Homeland Security

4. The Economy
Since the events of Sept. 11, 2001, we have learned how catastrophes can send
an economy reeling for many months. Numerous factors in the economy can affect
agricultural systems decisions. Managers in the manufacturing and processing areas
certainly must be in tune with changes in the economy. Some key factors of change
include oscillating inventory levels for supplies, available disposable income, new
housing starts, changes in gross domestic product (GDP), expansion/failures of
businesses, price levels, and changes in exports or imports.
Today, agriculture faces many changes in marketing channels for livestock,
increasing mergers of seed and chemical companies, and consolidation of equipment
suppliers. Yet new opportunities abound in the emergence of alternative fuel processors
and new crop initiatives.
AI3021- IT in Agriculture
5. Changing Societal Trends

Society’s attitudes are constantly changing and evolving. A number of attitude

changes have greatly impacted agricultural systems. Perhaps the most important has
been in the attitude toward the environment.
A growing spirit of conservation and preservation of wildlife and habitat has spurred
numerous changes in agricultural systems management. In the early 1990s, support for
the Pacific Northwest’s spotted owl changed many lives. The giant logging industry,
through public opinion, was forced to change its cultural and harvesting practices.
The industry continues, but not until after many systems changes were implemented.
Managers must look ahead and be considerate of society’s attitudes or perhaps pay a
larger price—being forced out of existence. Society now focuses more keenly on the
effects of production practices, including biotechnology, waste disposal, water quality,
chemical waste, and odor generation.
Other attitude changes during the past decade have included:
1. Change in attitude toward the use of electronics and
2. computing technologies
3. Change in attitude toward the use of credit
4. Change in attitude toward increased leisure time
5. Change in attitude toward the use of foreign products
6. Change in attitude toward health and Fitness
7. Change in attitude toward cultural diversity
8. Change in attitude toward higher education
9. Change in attitude toward human and animal health
10. Increased desire for food safety
11. Increased desire for protection from terrorism
Unfortunately, many of these attitude changes have brought about increased regulation
of a system manager’s activities. In order for the technical manager to be successful in
the long run, he or she must not limit his or her continuing education activities to
technical updates alone. One must develop a good
‘‘crystal ball’’ by following and participating in many cross educational
activities, including local, state, federal, and international politics. Some of these
changes in attitudes have encouraged whole ‘‘new’’ industries. The desire for increased
leisure and the desire to enjoy the ‘‘good life’’ have generated the new area of
agritourism. Industries such as hunting preserves, Bushing resorts, bed-and-breakfast
inns, grape vineyards and wineries, farmers markets, maple syrup festivals, and equine
events now abound.

6. The Competition
Someone else is always playing the same game. Decisions by the competition
sometimes affect the strategy of another ASM's planning and ultimately the outcome of
those decisions. Often, we may be affecting one and the same system. Awareness of
what others are doing to that system is important.
Competitors' decisions or actions can affect the economic well-being of other firms or
entities. The entry of new competitors does not always mean disaster for existing firms,
but it can—especially if management does nothing to secure its position. A new
competitor may force expansion in order to get lower cost per unit benefits. A new
competitor may force the changing of hours of operation, additional investments,
AI3021- IT in Agriculture
review of the product lines, expansion, or closures. Competition may come from foreign
markets, such as lower-cost produced pork from Argentina or Brazil.

The competition's addition of a new or improved product often changes the business or
management of a system. A seed company may add a new biotechnology-developed
seed. A steel building system might have improved life. The introduction of retort
packaging or irradiated food could change marketing, processing, or packaging
systems.
Competition might introduce new selling or marketing strategies. Promotion of Angus
beef might encourage growers to produce more Angus beef and less of other breeds of
cattle. New marketing strategies might change how one would package the product. Or
sometimes the new package itself creates the need for changes in processing.
Competition might obtain new customers, which gives them either economies of scale
or a new niche market. In the late 1980s the Florida corporation Naples Tomato Growers
landed the account to supply all tomatoes used by McDonald's Corporation. The
security of a contracted major account afforded them a number of new managerial
options.

New markets are always being sought. Competitors who might find them first win the
opportunity, at least briefly, to attain economies of scale. The adoption of NAFTA
opened many doors and opportunities for many agricultural industries. Clearly,
irrigation sales, food equipment sales, and aquaculture equipment sales to Mexico were
some of those new markets open for a short time. Perhaps one of the most exciting new
markets exists in China. Without a doubt China can be America's largest new market
for corn, soybeans, and other commodities.

The adoption of new, more efficient processes or management strategies can change
cost structure rapidly as well. In the 1970s Japan adopted new steel-making
technologies that forever changed the world industrial markets for steel. And in the
decade of 2000, the American technological expansion of Internet and software
technologies is greatly changing the face of all systems management.

Changes in Clients' Needs

The customer is always changing, and so are customers' needs. The ASM must develop
mechanisms that allow timely interaction as clients develop strategic plans. Regrettably,
many ASMs have become so engrossed in their own firm's plan that they simply
overlook the changing needs of the customer. And before long another firm is taking
care of those changing needs.

Consumer needs change constantly. A good case study is that of the consumer
acceptance of genetically modified products (GMs), or products from genetically
modified organisms (GMOs). European consumers and those in the United Kingdom
are keenly aware of food safety issues. Their perception (right or wrong) is that GMO-
developed food products may not be safe. Hence, they do not want GMO products
mixed with non-GMO products. ASMs in the United States may have to change many
operations, functions, and equipment pieces in order to have a food system that can
deliver a 100% non-GMO product.
AI3021- IT in Agriculture
And Americans have grown, literally. The average American is taller, larger, and
overweight. Clothes makers have to make those shirts bigger!
Changes in the ASM's or Firm's Purpose
The ASM and his or her firm is an entity that also changes, grows, adds and drops
obligations, and responds to new opportunities and challenges. The agricultural system
must change in personnel, functions, purpose, activities, and goals. A self-assessment
process is a vital part of ongoing management of the total system.

In summary, the performance of the system is greatly influenced by the ASM's

decisions. Success is determined by how well the ASM adjusts for changes in:

• Weather
• Seasonality
• Biological intrusions
• Technology
• Legal and political regulation
• The economy
• Societal trends

• Competition
• Customer or client needs
TOOLS FOR PLANNING, MANAGING, AND EVALUATING
The bottom line as to whether management is good or not is whether the system is
meeting performance specifications or improving. Validating that the system is on the
upswing can be very difficult. Verification requires measuring and analysis techniques.
To do this, a process requires a six-step loop:

1. Sensing
2. Information storage
3. Information processing and analysis
4. Evaluation
5. Decision
6. Action or change
7. And back to sensing, etc
A manager might utilize this testing process to improve the system on a weekly basis or
even a daily basis. Imagine the potato chip processing line manager we discussed earlier.
The chip process might be sampled and adjusted every few minutes or even seconds.
Improvement requires planning and management tools. Mathematical models are often
used for planning and evaluation. Any agricultural system can be analyzed utilizing
mathematical tools if it can be described logically. In most cases, approximation, record-
keeping, sensing devices, electronic monitors, satellite imaging, or statistical sampling
techniques will yield enough accurate data for management to begin planning and
evaluation.
Quite a number of today's best technical management tools are mathematical models
disguised as computer software packages. Many of those packages will be explored in
the upcoming chapters of this book. Good examples are spreadsheets, linear programs,
scheduling routines, and simulations.
AI3021- IT in Agriculture
One can begin to see that the skills required to be a successful ASM have a wide range,
based on the varied activities this technical manager must oversee. Managerial skills
are needed in:
• Accounting and finance

• Organizational planning
• Scheduling
• Systems reliability planning
• Personnel
• Human factors and human safety
• Environmental planning
• Pricing and costing
• Data information and management
• Transportation analysis
• Biological, chemical, and physical management
• Decision systems
• Systems integration
• Promotion and sales planning
• Legal and regulatory planning
Ideally, a strong ASM will have an academic preparation that is strong in engineering
technology, the sciences, and business. The use of statistics and computer tools is
paramount. Fortunately, there are many good software packages today to assist the
ASM. And the Internet can be an excellent source of current information. The upcoming
chapters will tie some of today's common agricultural systems challenges to a problem-
solving framework utilizing some of the more popular mathematical models and
contemporary software packages available today.

RELIABILITY OF AGRICULTURAL SYSTEMS

The goal of every agricultural systems manager (ASM) is to develop a total system that
functions without fail. We know that this is an unobtainable goal, yet we strive to
approach this by setting a goal of some successful percentage, such as 97% of the time
the system operates as we need it. In order to obtain a high reliability for a system we
must set a reasonable goal and plan for how to get there. The planning involves analysis
of each component and function of the system. Then we consider what devices are
replaced, which require backup, and how much we can afford to spend on our way to
attaining the system reliability goal.
Measuring performance of a system can be done in several ways, all depending on
which of the criteria we wish to evaluate. The technical manager spends a high
proportion of time evaluating the performance of the system. Every agricultural system
can be viewed as a group of processes and components (including humans) that must
AI3021- IT in Agriculture

perform satisfactorily and in a timely fashion to achieve the output we have specified.
Within our context we shall consider reliability to generally mean the probability that
our system performs successfully. And we as managers will predetermine what
"success" is.
We will now explore some of the management implications of our working definition
of reliability within the context of agricultural systems. In production management
reliability is most often viewed as the probability that the system will perform
satisfactorily when called upon under specified conditions. Thus, reliability of a system,
a subsystem, a function, or a component is measured in terms of probabilities.
Quantitatively, a component or system is expressed as 0.9999 or 0.94 or some positive
value less than 1.00. It should be noted that not all the system components necessarily
have to perform at the same time, but each component must operate at the proper time
for a sufficient period of time to ensure that the system accomplishes its purpose.
Operation of the jet propulsion engine on a space shuttle launch is a good example. The
engine must perform for only a few critical minutes in the proper way to be deemed
successful.
A critical part of the definition of reliability is the "specified conditions" under which a
unit or component is to function. Systems that would perform under all extremes of
heat, dust, humidity, poor field or manufacturing conditions, vibration, and
mismanagement and all possible conditions would be ideal—ideal but not very realistic.
Systems must be evaluated within the limits of their intended use. These conditions
should be explicitly stated or recognized before evaluating the performance reliability
of a system. For example, it would not be a fair evaluation to state that the reliability
was poor if a citrus harvest system failed because the hydraulic lift truck for moving
pallets was stolen, or if a computer designed to operate in an air-conditioned working
environment failed because it was exposed to 125°F temperatures in an incubation
chamber. It is the job of the technical or agricultural systems manager to assist in
planning and in setting needed, reasonable expectations.
Some critical processes may require component performances of extremely high
reliability (launching astronauts, embryo incubation, refrigeration, evaporators, etc.).
The words perform satisfactorily also carry great significance in the definition and in
understanding the performance reliability concept. Tolerances must be specified so as
to establish what acceptable reliability means. From the very beginning of the existence
of a system degradation begins. As each component within the system ages, wears, or
AI3021- IT in Agriculture

depletes with use, the output or system performance begins to vary. For a substantial
period, the variance may be so minimal that it is not worthy of notice. Eventually the
system may continue to function, but not at an acceptable level. A corn harvesting
system consisting of a combine, tractor, wagons, augers, bucket elevators, grain drier,
and storage bins is a good example. As the combine ages, field losses of grain increase.
As the augers and bucket elevators wear, grain kernel damage increases and system
flow rates decline. Drying rates may increase and fuel costs rise. There eventually
comes a time when the unacceptable level is reached, and the system is deemed no
longer reliable. One day the value may be 0.91 and the next day it can be deemed
inoperable (zero). The manager can elect to replace components, change the system, get
backup units, or even hire another system to perform the task.
Setting the acceptable levels of performance of various components is not always easy.
First, one must recognize what criteria are typical. The following criteria often need
acceptability levels established in field production agriculture:
• Speed of operation
• Crop loss acceptability
• Quality and condition of crop
• Timeliness of completion date or times
• Cost–profitability trade-offs
• Field efficiency
• Capacity of each system needed
To dwell on this topic of acceptability for just a moment, we might consider a farmer
evaluating the levels of acceptability regarding capacity of his soybean harvest system.
Due to the limited days of good weather in his Midwest fall season, he may know that
his harvest system must maintain the overall daily capacity of 55 acres per day. If
harvest capacity is less, he does not get all of the crops in before the snow arrives and
ends the season. Low reliability or high downtime, as he may phrase it, would not be
acceptable. Likewise, a corn producer must view the quality of the harvested crop. It
would do no good to continue running a combine if the threshing unit had become so
worn that the machine produced a crop with 20% of kernels cracked. The crop would
be evaluated poorly when graded at the grain terminal and deemed a very low market
grade (sample grade would yield a very low price)!
In biological processing, packaging, or food processing, added factors such as the
following could be determinants:
AI3021- IT in Agriculture

• Percent loss
• Yield
• Critical time durations met
• Cost per unit
• Food quality
• Food safety
• Contamination
• Cost of rework
Let's suppose a pecan processor is making bagged, crushed nuts for a candy
manufacturer. The pecan processor delivers several truckloads of nuts. It is found that
the nuts contain many pieces of shell. The candy manufacturer rejects the product (they
do not want lawsuits from customers who break their teeth). If it is found that the
screening devices are no longer capable of use because they are too worn, then the
reliability of the processing line becomes zero.
HUMAN INTERACTIONS
Most agricultural systems consist of mechanical equipment and humans. In some cases
biological, chemical, or physical processes are components as well. Humans are often
required for planning, initiation, maintenance, operation, vigilance, ending operations,
or any variety of tasks. They may provide the "backup" to any number of potentially
failed components.
If one considers only equipment or process factors in systems planning, then one is
assuming operator performance to have the probability of r = 1.00. Obviously, the
reliability of humans is not perfect, or 1.00. Leaving out the valuation of human
elements would give grossly inflated systems reliabilities, as often happens. However,
the proper management of human interaction can lead toward exceptionally high
systems reliability, as we will see later in this chapter.
Humans are much more complex than any machine or process used in agricultural
systems today. The challenge of duplicating higher human functions such as perception,
recognition, and decision making has just begun, through artificial intelligence
algorithms and electronic circuitry. The field of robotics is still expensive and in its
infancy.
Human limitations are numerous. They are less stable than machines and are influenced
by and more responsive to the work environment. Human performance is affected by
physiological conditions, fatigue, noise, incentives, rewards, previous learning, and
AI3021- IT in Agriculture

conditioning (good and bad). However, thanks to the study of human factors, or
ergonomics, it is possible to treat human operators mathematically, as one does for other
components and processes as we estimate reliability of systems. In terms of inputs and
outputs, a common descriptive language exists from empirically derived research. This
permits a mathematical treatment that can be applied to man, machine, or processes.
We will return to this topic.
ESTIMATING THE SYSTEMS RELIABILITY VALUES
There are many questions for which estimation of reliability can assist in providing
technical management answers. Always remember: The system must work when you
need it. If it doesn't, all other aspects of a system are irrelevant. Here are just a few of
the decisions that quantification can help answer:
• What percent of the time does this system really work?
• Which machine should be replaced to gain reliability?
• Should a new machine (or person) be purchased, or can a used machine be
utilized?
• If other machines are available for "backup," how will this affect performance?
• How many "backup" units are needed?
• Which is more profitable renting, buying, or leasing?
• Will the system work if a particular unit fails?
• How does the human operator or manager in the system affect the probability
of success? Do we need more management or less?
• How much will I gain in efficiency and capacity by increasing reliability or
decreasing it and saving costs?
• What level of reliability is economically acceptable?
Production managers, salespersons, service personnel, and design engineers need to
fully understand that system reliability is inherently the absolute bottom line in selecting
equipment, people, and resources. Granted, selections are often made for other reasons
(economic, safety, etc.), but reliability must always be at an acceptable level.
Estimating System Reliabilities
Components in Series
Many agricultural systems are arranged in series. The successful operation of a series
system depends upon the successful performance of each component in the system:
man; machine; or process. Two conditions often exist:
(1) Failure of any given unit results in a complete system failure, and
AI3021- IT in Agriculture

(2) the component failures are independent of each other.

Gordon showed that in series systems the probability that a system operates acceptably
is the product of the reliabilities of the individual units or components.
If, for example, there are three components in a system, each with a reliability of
0.90, the reliability of the system would be a product of the three, or 0.729. Lusser
presented the original formula for sequential events as:

The inherent weakness of most agricultural systems is the sequential nature, leading to
overall low system performance reliability [4]. Think of a simple wheat harvesting
system consisting of a tractor (0.90) to pull a wagon (0.90) that the combine (0.85)
unloads into a storage tank (Fig. 2-1). The system reliability is only 0.6885. One unit
fails and the whole system stops.
As more units are added to an agricultural system, each unit must be very close to unity
(1.00) if the system is to remain acceptable. There are three possibilities for improving
a system:
1. Replace components of lowest reliability with higher-valued units.
2. Use only components of high reliability (new equipment).
3. Use redundant components (provide dedicated "backup" units).

With parallel units, there are two or more that are performing the same function or are
available to perform the same function at any particular time. This is referred to as either
"backup" or redundancy. NASA missile research showed that parallel system reliability
of like units can be estimated by combining the probabilities of unit success
(reliabilities) of the individual units using the following formula [2]:

where:
• Rsystem is the reliability of the system.
• r is the reliability of a single unit.
AI3021- IT in Agriculture

• m is the number of components in parallel for each function.

• n is the number of functions the unit must perform.
Suppose one were baling hay and a tractor is needed to pull and power the baler. The
farmer needs only a single functioning tractor at one time to pull the hay baler. However,
if he had two identical tractors, both available 100% of the time if needed and, say, with
unit reliabilities of 0.90, the joint probability of having a tractor available to pull the
baler would be (Fig. 2-2):

Thus, even with relatively low component reliabilities such as 0.70, a system with four
units in parallel could achieve a system reliability of 0.992. In some systems the
components could be very different types, such as a human backing up a machine unit,
or vice versa.
Achieving redundancy, or "backup," in a system can be accomplished in several ways.
Some common management alternatives are:
• Purchasing another machine
• Borrowing a "backup" machine when needed
• Using a unit from another operation
• Leasing another unit
• Renting another unit
• Assured availability warranty from manufacturer

Equation (2-2) is very useful because it expresses the real world of agricultural systems
fairly well. The equation, however, applies only where the components in parallel for
each of the n functions is exactly the same and unit reliability of each component is the
same. Where these conditions do not apply, the derivation becomes more complex. And
in reality it is rare to find two components or processes that are identical. For example,
AI3021- IT in Agriculture

a new tractor may be in use with a reliability value of 0.95; but if the new tractor were
to fail, the manager might bring in the old tractor from the shed (0.78) to back the newer
unit up. This situation is very typical in field production agriculture and processing.
Complexity occurs in deriving a set of equations, since the selection of the first unit is
a parameter open to management. An iterative technique is essential.
Heterogeneous Units
Redundant units in parallel possessing different reliability values can be termed
heterogeneous backup units. An iterative technique would work as follows: For the
components in parallel, select the first. If the system only had that one, that subsystem
reliability, Rs, would be equal to that of the component, R1. The reliability of the
subsystem that included both components 1 and 2 (call it Rs(1,2)) would be the
probability that the first unit functions (R1) plus the probability that the first unit fails
(1 - R1) and the backup unit functions (R2). These last two reliabilities are multiplied
to get the reliability of the backup system's functioning. Therefore, the reliability of the
subsystem with a heterogeneous backup unit is

Suppose two tractors are in parallel to form some subsystem.

Let’s say that tractor 1 has a unit reliability of 0.90 and
unit 2 is 0.78 (Fig. 2-3). Using Eq. (3), the reliability calculation
would be:

The dynamics of production agriculture systems can become quite complex. Tractors
may back up several subsystems, performing several functions, or be unable to serve as
backups because of incompatibility of components or unmatched horsepower
requirements. Reliabilities seldom are identical. The best approach is to calculate
AI3021- IT in Agriculture
reliabilities for each subsystem separately and then to combine the subsystem values to
attain the complete system reliability.
Suppose farmer A had two enterprises: dairy and peanuts. To some extent, the two
separate agricultural systems must share equipment, such as tractors. Farmer A must
recognize the strengths and weaknesses of this arrangement. Let's analyze farmer A's
peanut harvesting operation, as shown in the following table and Figure 2-4. Suppose it
consists of a large tractor pulling a peanut combine. A wagon attached to the combine
receives the peanuts. Another small tractor pulls the wagon away, empties the load and
returns just in time to exchange wagons with the combine unit. The iterative process
assumes that the unit of highest reliability is always used first. This is the usual
management situation.
The calculations of each subsystem would be as follows:
Subsystem 1:
• Rss1 = 0.83 + 0.17 * (0.83)^3 = 0.9272
Subsystem 2:
• Rs(1,2) = 0.85 + (1 - 0.85) * 0.72 = 0.9580
Farmer A's System

Subsystem Machinery available Reliabilities

1 1 100-hp tractor dedicated to pulling combine 0.83 each

100-hp tractor backs up pulling unit, small

1 tractor in subsystem 5, and tractor working at the
dairy operation

2 1 peanut combine 0.85

2 1 backup combine 0.72

3 1 wagon behind combine 0.90

4 1 wagon behind tractor 0.90

1 extra wagon to back up either subsystem 3 or

4 0.90
4

5 1 40-hp tractor to pull wagon

6 1 40-hp tractor to pull wagon

Note: The table provides information about the machinery available for each
subsystem, along with their respective reliabilities.
AI3021- IT in Agriculture
AI3021- IT in Agriculture

Estimating Values of Individual Components

Just how does one attain the reliability values for machines, components, and functions?
Absolutely the best way is to keep records on these units. Fortunately, computers and
spreadsheet software make this task much easier today. Line foremen can record
downtimes, as can mechanics or service managers. In the case of agricultural field
machines, this is certainly "doable." One of the world's largest field production
sugarcane growers actually tracks each field machine and keeps a life record via
computer database. A farmer could keep uptime and downtime records on a
spreadsheet.
In the case of processing plants, maintenance records often exist, and data estimates can
be made by "recouping" the past information on items such as blenders, mixers,
conveyors, bagging machines, and chemical processes. Where no life records exist, a
good manager can collect sample data using good statistical techniques. Data can
sometimes be gathered from other plants or the engineering firms producing the
devices.
The literature in processing journals does contain some reliability data. And when all
else fails, an ASM could interview users of the machines and processes to backtrack in
time to find failures and downtimes. The real bottom line is that attaining accurate unit
reliability data requires forward planning. However, the effort to attain the data will
yield great rewards in system performance.
Estimating Human Component Values
Many situational factors affect human performance. Operator unit reliabilities can range
from zero to 0.99999 reliability. It becomes quite difficult to develop generalized
relationships. The trade-off considerations that can be applied are consequently
AI3021- IT in Agriculture
qualitative. Costs, hazards, state of the technology, and other factors often influence the
human role in an agricultural system. It was also shown that the reliability in a space
system with a maintenance person available could be higher than that of the same
system with automated devices as a monitor or "fail-safe" unit. Humans can seldom be
rivaled by equipment. The following illustrates the trade-offs that exist with the use of
humans.
Human long-term unit values seldom exceed 0.78. This is because there are so many
time deductions one must make. Ignoring time lost for weekends and hours beyond 40
per week, the following lost time applies:
• Sick time
• Late time
• Vacation time
• Family leaves
• Strike time
• Break times
• Other
Humans are also subject to errors from repetitive tasks. The value assigned to human
subjects varies greatly, depending on how they are inserted into the system. With proper
managerial planning and backups, the values can be very high.
Let's take a closer look at the true value of a human in an agricultural system. Suppose
a woman operates a bagging machine in a line operation that produces bags of garden
mulch for the Super Duper market outlets. In a year, if she worked an eight-hour day,
five days a week, she would need to be available 2080 hours per year. But she is not
likely to be able to deliver this. Consider the following time losses:

The first reaction would be that this is unacceptable. But, this is just being human. It
would be easy to develop a much worse scenario for an employee. So how does the
plant manager cope with labor on a line that might involve 20 or more employees?
First, let's assume that the manager could back up this employee with another available
employee. The manager might call in another qualified 1 individual off-shift. Or maybe
there is a pool of employees available. The calculation would 2 be:
AI3021- IT in Agriculture

Managerial Implications
The use of parallel or redundant units becomes a very important factor in production
management decisions, such as deciding what unit to replace and whether to buy new
or used equipment. Reliability and machinery labor costs are clearly traded off to attain
some acceptable level of performance.
It becomes clear that the use of redundant subsystems is often more economical than
simply purchasing new units with higher reliability. Comparisons between ownership
and operating costs of new purchases versus several older units and their maintenance
and repair operating costs are necessary. Leasing and rental units must also be
considered. Individuals or corporations "starting out" or with limited capital obviously
have some alternatives that may be quite viable.
The human operator can be used to great advantage. If inserted properly into a system,
reliability can be increased several-fold to save capital outlay.
Weighing the Cost of Attaining System Reliability
One can come to the supposition very quickly that most systems must achieve a 0.95
total reliability or better to be considered "successful." There are almost always
alternatives. And each alternative has its costs. The decision could be driven by how to
get an acceptable system reliability for the lowest cost. Let's consider a struggling
college graduate, Bob, desiring to farm. He owns an older combine of unit reliability of
0.88. He knows this is not good enough. So what are his alternatives?
Suppose his dad is willing to allow Bob to use his new combine of 0.96 reliability as a
backup, but only when he is not combining himself. The calculation would be:
AI3021- IT in Agriculture

In a processing situation, some devices might be extremely expensive, such as a reactor

in an ethanol processing plant. The backup alternative may be maintaining an extensive
parts inventory or even an emergency contract with an engineering firm to perform
immediate services.
Another alternative is to introduce the concept of scheduled repair versus maintenance.
Suppose the huge sugarcane harvest system has hundreds of tractors. Based on past
records, the managers may know what breaks down and when. They might actually shut
down the operation of units with higher hours and rebuild the transmissions, hydraulics,
and engines before they break down. This might be expensive and require a larger pool
of tractors, but it might be cheaper than buying new units or stocking even more backup
units.
This concept is often used in situations where downtime is either extremely expensive
or critical (no one wants their army tank to break down during battle)! Or the grain
elevator operation does not want the bucket elevator to break down during a key harvest
period. In any case, the ASM needs to be a true thinker!
This type of evaluation and trade-off consideration should be ongoing. Arguably,
reliability planning could be the most important agricultural systems decision—but the
one most often neglected.
AI3021- IT in Agriculture

Simulation of Crop Growth and Field Operations

Definition: Crop growth simulation models are mathematical representations of the

complex interactions between plants, the environment, and management practices. They
simulate the growth and development of crops under various conditions, providing
insights into potential yields, resource requirements, and the impact of different
management strategies.
Key Components:
• Plant Growth Models: These models describe the physiological processes of
plants, such as photosynthesis, respiration, and nutrient uptake. They incorporate
factors like leaf area development, light interception, and dry matter
accumulation.
o Example: The CERES-Wheat model simulates the growth of wheat by
considering factors like temperature, water availability, and nitrogen
uptake.
• Environmental Models: These models simulate factors like temperature,
precipitation, sunlight, and soil conditions that affect crop growth. They can be
based on historical data, weather forecasts, or climate change projections.
o Example: The EPIC (Environmental Policy Integrated Climate) model
simulates the impact of climate change on crop yields in different regions.
• Management Practices: These models incorporate the effects of various
agricultural practices, such as irrigation, fertilization, and pest control. They can
be used to evaluate the impact of different management strategies on crop
performance.
o Example: The APSIM (Agricultural Production Systems Simulator)
model can simulate the effects of different irrigation regimes on crop
yields and water use efficiency.
Benefits:
• Predicting Yields: Simulate crop growth under different scenarios to estimate
potential yields and identify the optimal management practices to maximize
production.
• Optimizing Resource Use: Identify the optimal allocation of resources (water,
fertilizer, pesticides) to maximize yields while minimizing costs and
environmental impact.
AI3021- IT in Agriculture

• Evaluating Management Strategies: Assess the impact of different

management practices on crop performance and sustainability. For example,
simulate the effects of different tillage practices on soil erosion and crop yields.
• Risk Assessment: Evaluate the risks associated with climate variability and
other uncertainties. For example, simulate the impact of drought or extreme
weather events on crop production.
Example: A farmer is considering using a new irrigation system to improve crop yields.
By using a crop growth simulation model, the farmer can simulate the impact of
different irrigation regimes on crop growth and water use efficiency, helping them make
an informed decision about whether to invest in the new system.
Growing degree-days (GDDs) are a measure of heat accumulation used to predict plant
and pest development rates. They are calculated by adding up the daily average
temperature (above a certain base temperature) over a period of time. GDDs can be used
to estimate the date that a crop will reach maturity, when a pest will emerge from
dormancy, or when a flower will bloom.
Here's a breakdown of the concept:
• Base temperature: This is the minimum temperature at which a plant or pest
can begin to grow or develop. It varies among different species and can be
affected by factors such as genetics and environmental conditions.
• Daily average temperature: This is the average of the maximum and
minimum temperatures for a given day.
• GDD calculation: GDDs are calculated by subtracting the base temperature
from the daily average temperature and adding up the values for each day. If the
daily average temperature is below the base temperature, the GDD value is zero.
Here's an example of how to calculate GDDs:
• Base temperature for corn: 50°F
• Daily maximum temperature: 85°F
• Daily minimum temperature: 60°F
• Daily average temperature: (85°F + 60°F) / 2 = 72.5°F
• GDDs for that day: 72.5°F - 50°F = 22.5 GDDs
How GDDs are used:
• Predicting crop maturity: By tracking the accumulation of GDDs over the
growing season, farmers can estimate when their crops will reach maturity and
be ready for harvest.
AI3021- IT in Agriculture

• Timing pest control: GDDs can be used to predict the emergence of pests and
time pesticide applications accordingly.
• Selecting varieties: GDDs can help farmers select varieties of crops that are
well-suited to their local climate and growing season.
• Monitoring plant health: GDDs can be used to monitor the progress of plant
growth and identify potential problems.
Limitations:
• GDDs are a simplified model of plant growth and do not account for all factors
that can affect plant development.
• GDDs may not be accurate for plants that are stressed by environmental
conditions or other factors.
• GDDs may vary depending on the specific location and microclimate.
Overall, GDDs are a valuable tool for farmers and agricultural researchers. By
understanding how to calculate and use GDDs, farmers can make more informed
decisions about crop management and improve their yields.

Optimizing the Use of Resources

Definition: Resource optimization involves allocating available resources (e.g., land,

water, labor, capital) in the most efficient way to achieve desired objectives, such as
maximizing profit or minimizing costs.
Key Techniques:
• Linear Programming: A mathematical optimization technique used to
allocate limited resources to maximize or minimize a linear objective function
subject to linear constraints. It is suitable for problems with a single objective
and linear relationships between variables.
o Example: A farmer wants to maximize profit by planting two crops,
corn and soybeans. The farmer has 100 acres of land available and a
limited budget for fertilizer. The profit per acre for corn is $100, and for
soybeans is $80. The fertilizer requirements for corn are 2 units per acre,
and for soybeans are 1 unit per acre. The total available fertilizer is 150
AI3021- IT in Agriculture

units. The farmer can use linear programming to determine the optimal
allocation of land between corn and soybeans to maximize profit.
• Nonlinear Programming: A more general optimization technique that can
handle nonlinear relationships between variables. It is suitable for problems with
multiple objectives or nonlinear constraints.
o Example: A farmer wants to minimize the cost of irrigation while
maintaining a certain level of crop yield. The cost of irrigation may be a
nonlinear function of the amount of water applied, and the crop yield may
also be a nonlinear function of water availability. Nonlinear programming
can be used to find the optimal irrigation schedule.
• Dynamic Programming: A technique for solving optimization problems that
involve sequential decisions. It is suitable for problems where decisions made at
one time period affect the outcomes of decisions made in subsequent periods.
o Example: A farmer wants to determine the optimal timing of fertilizer
applications throughout the growing season to maximize crop yield.
Dynamic programming can be used to consider the impact of fertilizer
applications at different stages of crop growth.
Applications:
• Crop Allocation: Determining the optimal mix of crops to grow on a given
piece of land, considering factors such as soil type, climate, market demand, and
profitability.
• Irrigation Scheduling: Optimizing irrigation water use to maximize crop
yields while minimizing water wastage and environmental impact.
• Fertilizer Management: Determining the optimal amount and timing of
fertilizer application to maximize crop yields and minimize nutrient losses.
• Machinery Allocation: Assigning machinery to tasks in a way that maximizes
efficiency and minimizes costs.
• Farm Management: Optimizing the overall management of a farm to
maximize profitability and sustainability.

Linear Programming

Definition: Linear programming is a mathematical optimization technique used to find

the optimal solution to a problem with a linear objective function and linear constraints.
AI3021- IT in Agriculture

Basic Components:
• Objective Function: The function to be maximized or minimized.
• Decision Variables: The variables that can be adjusted to optimize the
objective function.
• Constraints: The limitations or restrictions on the decision variables.
Example:
• Problem: A farmer wants to maximize profit by planting two crops, corn and
soybeans. The farmer has 100 acres of land available and a limited budget for
fertilizer. The profit per acre for corn is $100, and for soybeans is $80. The
fertilizer requirements for corn are 2 units per acre, and for soybeans are 1 unit
per acre. The total available fertilizer is 150 units.
• Linear Program:
o Maximize Profit = 100Corn + 80Soybeans
o Subject to:
▪ Corn + Soybeans <= 100 (land constraint)
▪ 2*Corn + Soybeans <= 150 (fertilizer constraint)
▪ Corn, Soybeans >= 0 (non-negativity constraint)
Solution:
The optimal solution is to plant 50 acres of corn and 50 acres of soybeans, which will
result in a maximum profit of $9000.

Project Scheduling

Definition: Project scheduling involves planning, organizing, and coordinating tasks to

ensure that a project is completed on time and within budget.
Key Techniques:
• Gantt Charts: Visual representations of project timelines, showing the
duration of each task and their dependencies. They are useful for tracking project
progress and identifying potential bottlenecks.
o Example: A Gantt chart for a construction project would show the
duration of each phase of the project, such as site preparation, foundation
work, framing, and finishing.
AI3021- IT in Agriculture

• Network Diagrams: Graphical representations of project activities and their

relationships, often using the Critical Path Method (CPM) to identify the critical
path (the sequence of tasks that determine the project's overall duration).
o Example: A network diagram for a crop production project would show
the sequence of tasks, such as planting, irrigation, fertilization, and
harvesting. The critical path would be the longest sequence of tasks that
must be completed before the project can be finished.
• PERT (Program Evaluation and Review Technique): A probabilistic
network analysis technique that considers the uncertainty in task durations. It is
useful for projects with uncertain task times.
o Example: A PERT chart for a research project would show the
estimated duration of each task, along with the optimistic, most likely,
and pessimistic estimates.
Applications:
• Construction Projects: Planning and managing the construction of buildings,
infrastructure, or other projects.
• Agricultural Operations: Scheduling planting, harvesting, and other
agricultural activities.
• Research Projects: Organizing and coordinating research tasks to ensure
timely completion.

Artificial Intelligence and Decision Support Systems

Definition: Artificial intelligence (AI) refers to the development of intelligent agents

that can reason, learn, and solve problems. Decision support systems (DSS) are
computer-based systems that help decision-makers analyze data and make informed
decisions.
Applications in Agriculture:
• Crop Yield Prediction: Using AI algorithms to predict crop yields based on
historical data, weather forecasts, and other factors. This can help farmers make
better decisions about planting, harvesting, and marketing.
• Pest and Disease Management: Employing AI to detect and identify pests and
diseases early, enabling timely intervention to prevent crop damage.
AI3021- IT in Agriculture

• Precision Agriculture: Using AI-powered technologies (e.g., drones, sensors)

to collect data on crop health, soil conditions, and other factors to optimize
resource management. This can help farmers use resources more efficiently and
reduce their environmental impact.
• Decision Support Systems: Developing DSS to help farmers make informed
decisions about planting, harvesting, marketing, and other agricultural activities.
DSS can provide farmers with access to relevant data, analytics tools, and expert
advice.
Key AI Techniques:
• Machine Learning: Algorithms that allow computers to learn from data and
improve their performance over time. For example, machine learning can be used
to develop models that predict crop yields based on historical data.
• Natural Language Processing: Enabling computers to understand and
respond to human language. This can be used to develop chatbots that can answer
farmers' questions about agricultural practices.
• Computer Vision: Algorithms that allow computers to interpret and
understand visual information. This can be used to develop systems that can
automatically identify weeds or detect plant diseases from images.
Example: A farmer can use a DSS to analyze weather data, soil information, and crop
prices to determine the optimal planting date for a particular crop. The DSS can also
provide recommendations for fertilizer application, irrigation scheduling, and pest
control.

Artificial Intelligence:
Benefits of AI in agriculture
Until recently, using the words AI and agriculture in the same sentence may have
seemed like a strange combination. After all, agriculture has been the backbone of
human civilization for millennia, providing sustenance as well as contributing to
economic development, while even the most primitive AI only emerged several decades
ago. Nevertheless, innovative ideas are being introduced in every industry, and
agriculture is no exception. In recent years, the world has witnessed rapid advancements
in agricultural technology, revolutionizing farming practices. These innovations are
becoming increasingly essential as global challenges such as climate change, population
growth together with resource scarcity threaten the sustainability of our food system.
Introducing AI solves many challenges and helps to diminish many disadvantages of
traditional farming.
AI3021- IT in Agriculture
Data-based decisions
The modern world is all about data. Organizations in the agricultural sector use data to
obtain meticulous insights into every detail of the farming process, from understanding
each acre of a field to monitoring the entire produce supply chain to gaining deep inputs
on yields generation process. AI-powered predictive analytics is already paving the way
into agribusinesses. Farmers can gather, then process more data in less time with AI.
Additionally, AI can analyze market demand, forecast prices as well as determine
optimal times for sowing and harvesting.
Artificial intelligence in agriculture can help explore the soil health to collect insights,
monitor weather conditions, and recommend the application of fertilizer and
pesticides. Farm management software boosts production together with profitability,
enabling farmers to make better decisions at every stage of the crop cultivation process.
Cost savings
Improving farm yields is a constant goal for farmers. Combined with AI, precision
agriculture can help farmers grow more crops with fewer resources. AI in farming
combines the best soil management practices, variable rate technology, and the most
effective data management practices to maximize yields while minimizing minimize
spending.
Application of AI in agriculture provides farmers with real-time crop insights, helping
them to identify which areas need irrigation, fertilization, or pesticide treatment.
Innovative farming practices such as vertical agriculture can also increase food
production while minimizing resource usage. Resulting in reduced use of herbicides,
better harvest quality, higher profits alongside significant cost savings.
Automation impact
Agricultural work is hard, so labor shortages are nothing new. Thankfully, automation
provides a solution without the need to hire more people. While mechanization
transformed agricultural activities that demanded super-human sweat and draft animal
labor into jobs that took just a few hours, a new wave of digital automation is once more
revolutionizing the sector.
Automated farm machinery like driverless tractors, smart irrigation, fertilization
systems, IoT-powered agricultural drones, smart spraying, vertical farming software,
and AI-based greenhouse robots for harvesting are just some examples. Compared with
any human farm worker, AI-driven tools are far more efficient and accurate.
Applications of artificial intelligence in agriculture
Traditional farming involves various manual processes. Implementing AI models can
have many advantages in this respect. By complementing already adopted technologies,
an intelligent agriculture system can facilitate many tasks. AI can collect and process big
data, while determining and initiating the best course of action. Here are some common
use cases for AI in agriculture:
AI3021- IT in Agriculture
Optimizing automated irrigation systems
AI algorithms enable autonomous crop management. When combined with IoT
(Internet of Things) sensors that monitor soil moisture levels and weather conditions,
algorithms can decide in real-time how much water to provide to crops. An autonomous
crop irrigation system is designed to conserve water while promoting sustainable
agriculture and farming practices. AI in smart greenhouses optimizes plant growth by
automatically adjusting temperature, humidity, and light levels based on real-time data.

Detecting leaks or damage to irrigation systems

AI plays a crucial role in detecting leaks in irrigation systems. By analyzing data,
algorithms can identify patterns and anomalies that indicate potential leaks. Machine
learning (ML) models can be trained to recognize specific signatures of leaks, such as
changes in water flow or pressure. Real-time monitoring and analysis enable early
detection, preventing water waste together with potential crop damage.
AI also incorporates weather data alongside crop water requirements to identify areas
with excessive water usage. By automating leak detection and providing alerts, AI
technology enhances water efficiency helping farmers conserve resources.
Crop and soil monitoring
The wrong combination of nutrients in soil can seriously affect the health and growth
of crops. Identifying these nutrients and determining their effects on crop yield with AI
allows farmers to easily make the necessary adjustments.
While human observation is limited in its accuracy, computer vision models can monitor
soil conditions to gather accurate data necessary for combatting crop diseases. This
plant science data is then used to determine crop health, predict yields while flagging
AI3021- IT in Agriculture
any particular issues. Plants start AI systems through sensors that detect their growth
conditions, triggering automated adjustments to the environment.
In practice, AI in agriculture and farming has been able to accurately track the stages of
wheat growth and the ripeness of tomatoes with a degree of speed and accuracy no
human can match.

Detecting disease and pests

As well as detecting soil quality and crop growth, computer vision can detect the
presence of pests or diseases. This works by using AI in agriculture projects to scan
images to find mold, rot, insects, or other threats to crop health. In conjunction with
alert systems, this helps farmers to act quickly in order to exterminate pests or isolate
crops to prevent the spread of disease.
AI technology in agriculture has been used to detect apple black rot with an accuracy
of over 90%. It can also identify insects like flies, bees, moths, etc., with the same degree
of accuracy. However, researchers first needed to collect images of these insects to have
the necessary size of the training data set to train the algorithm with.
Monitoring livestock health
It may seem easier to detect health problems in livestock than in crops, in fact, it’s
particularly challenging. Thankfully, AI for farming can help with this. For example, a
company called CattleEye has developed a solution that uses drones, cameras together
with computer vision to monitor cattle health remotely. It detects atypical cattle
behavior and identifies activities such as birthing.
CattleEye uses AI and ML solutions to determine the impact of diet alongside
environmental conditions on livestock and provide valuable insights. This knowledge
can help farmers improve the well-being of cattle to increase milk production.
AI3021- IT in Agriculture

Intelligent pesticide application

By now, farmers are well aware that the application of pesticides is ripe for optimization.
Unfortunately, both manual and automated application processes have notable
limitations. Applying pesticides manually offers increased precision in targeting
specific areas, though it might be slow and difficult work. Automated pesticide spraying
is quicker and less labor-intensive, but often lacks accuracy leading to environment
contamination.
AI-powered drones provide the best advantages of each approach while avoiding their
drawbacks. Drones use computer vision to determine the amount of pesticide to be
sprayed on each area. While still in infancy, this technology is rapidly becoming more
precise.

Yield mapping and predictive analytics

Yield mapping uses ML algorithms to analyze large datasets in real time. This helps
farmers understand the patterns and characteristics of their crops, allowing for better
AI3021- IT in Agriculture
planning. By combining techniques like 3D mapping, data from sensors and drones,
farmers can predict soil yields for specific crops. Data is collected on multiple drone
flights, enabling increasingly precise analysis with the use of algorithms.
These methods permit the accurate prediction of future yields for specific crops, helping
farmers know where and when to sow seeds as well as how to allocate resources for the
best return on investment.
Automatic weeding and harvesting
Similar to how computer vision can detect pests and diseases, it can also be used to
detect weeds and invasive plant species. When combined with machine learning,
computer vision analyzes the size, shape, and color of leaves to distinguish weeds from
crops. Such solutions can be used to program robots that carry out robotic process
automation (RPA) tasks, such as automatic weeding. In fact, such a robot has already
been used effectively. As these technologies become more accessible, both weeding and
harvesting crops could be carried out entirely by smart bots.
Sorting harvested produce
AI is not only useful for identifying potential issues with crops while they’re growing.
It also has a role to play after produce has been harvested. Most sorting processes are
traditionally carried out manually however AI can sort produce more accurately.
Computer vision can detect pests as well as disease in harvested crops. What’s more, it
can grade produce based on its shape, size, and color. This enables farmers to quickly
separate produce into categories — for example, to sell to different customers at
different prices. In comparison, traditional manual sorting methods can be painstakingly
labor-intensive.

Surveillance
Security is an important part of farm management. Farms are common targets for
burglars, as it’s hard for farmers to monitor their fields around the clock. Animals are
AI3021- IT in Agriculture
another threat — whether it’s foxes breaking into the chicken coop or a farmer’s own
livestock damaging crops or equipment. When combined with video surveillance
systems, computer vision and ML can quickly identify security breaches. Some systems
are even advanced enough to distinguish employees from unauthorized visitors.

AI and its Application

The concept of expert systems, artificial intelligence, ‘‘fuzzy logic,’’ and
knowledge-based decision support systems (information systems) was
accepted rapidly in business and industry. Production control systems in
factories have included these concepts so that operating problems, such as bad
welds, can be either prevented or fixed much more quickly than under simpler
control systems.

Robotics
Robotics has been considered a part of the overall field of artificial
intelligence by many authors. A robot is not just an automatic welding
machine on a car body production line, but a machine that can be programmed
to do a variety of tasks and that can interact with its environment. A robot
may need to make a decision about whether an object that it must select is a
nut or a bolt, a green tomato or a ripe, red tomato. Thus robots need to ‘‘see,’’
recognize objects, and make decisions, so they need intelligence ‘artifi cial’’
intelligence.

Natural Language
For easier and greater application of the computer in our working
world, we need better and more natural ways for hu- mans to communicate
with the computer. The computer needs to be shown more about how we
communicate so that we can spend less time learning how the computer
communicates. This is a difficult field, but progress is being made. The
programs that can process natural language sentences must determine which
words are the noun, the verb, and the object, In addition, voice recognition by
the computer is a part of this, because we often do not want to or are unable to
type our input, but should speak to the computer. The development of
programs that understand human natural language is a vital part of the whole
artificial intelligence investigation, because this requires intelligence. We are
going no further into this area.

Fuzzy Logic
To simulate human intelligence, artificial intelligence must be able to handle
more than just numbers, yet the heart of any computer is just a very fast
processor of data in binary form, zeros and ones. Our programming
languages, such as FORTRAN, C++, and Java, have allowed us to program
computers with English words and mathematical symbols, yet most
programs are very precise and objective in their results. They usually
produce a table of numbers, and they can rank these results numerically, even
AI3021- IT in Agriculture
with huge numbers of possibilities.
But humans often must make decisions in which the input data and the
resulting output are not in neat mathematical or completely logical form. For
example, a computer using a linear programming algorithm can select the one
best combination of several ingredients that will result in the lowest-cost feed
mixture that meets the specifications in terms of protein, fat, fiber, and total
digestible nutrients from among an almost infinite number of possible
combinations. However, until artificial intelligence, the computer could not
solve less precise problems, such as determining the most profitable
combination of crops to plant for the coming year, with prices and costs being
uncertain. And managers often want a recommendation in terms somewhat
like a weather forecast, such as, ‘‘If you do not spray, chances are about 1 in
5 that the disease will cause more damage than the spraying cost.’’
Expert Systems
An expert system simulates a human expert in a narrow subject matter
domain. For example, to develop an expert system program to give soybean
growers advice about insect problems like a consulting entomologist might give,
the program needs to do the following:

1. Ask the grower some general questions to find out what the problem
is.
2. Find out the farm’s location, variety of soybeans, stage of growth,
identification, or description of the bugs.
3. Finally, give the best estimate the system can on the type of insect,
whether or not treatment is needed at this time, and when and what
type of insecticide is recommended if treatment is needed.

Programs such as these probably will not replace the consultant, but they
will be used by the consultant to give better advice and probably to produce
good printed reports for the client. Many agricultural expert systems are
diagnostic in nature, such as the example just given. In addition, expert
systems may be (and a few have been) developed for making technical
management decisions, such as purchase of new equipment, deciding on a
crop rotation, making a marketing plan for grain, or culling livestock from a
breeding herd.
AI3021- IT in Agriculture
Knowledge-Based Decision Support Systems
The term knowledge-based decision support systems (or decision support systems)
includes expert systems but covers a broader range of program types .Often it
would be advantageous. for an expert system to have available the latest
data from the commodity futures market or cash markets or to be able to
run a simulation program with current weather data to get crop yield
estimates in order to give the user the best up-to- date expert advice. When an
expert system uses a database, a spreadsheet, or some other external
program, such as a simulation program, the whole integrated system is
referred to as a decision support system. An expert system is also a knowledge-
based decision system or decision support system, because it contains the
knowledge of the expert and it helps, or supports, the user in making a
decision.
The word support emphasizes the important idea that the computer is not
controlling the decision. It is not the decision maker. It helps (supports) the
human decision maker by keeping track of many factors, whereas the
decision maker is also likely taking into account other factors, especially
more subjective ones. A citrus production manager in Florida makes a very
sound point about the use of computers in management: ‘‘Use the computer
for what it can do best, calculating and remembering lots of data, and use the
human for what the person can do best, integrating the output of the decision
sup- port system and other factors, including the human’s experience, and
come to a decision based on all these things.’’
More recently, a broader common term has come to include decision support
systems: information systems; the field has become known as information
technology (IT).
Models of Decision Support system
AI3021- IT in Agriculture
Eg: Decision support system for crop disease.

Decision Support System Applications

Potential applications are innumerable, but many of them will
fi t under one of the following types.
• Diagnostics
• Marketing
• Systems operations management
• Automatic control
• Strategic planning
Static and Dynamic Decision Support Systems
Especially for agricultural systems, the idea of dynamic vs. static decision
support systems is important for the development and operation of these
systems. Static systems are those that are used, usually, at a single point in
time; and each use is a new one, keeping no information from the last time it
was used or ‘‘run.’’ A diagnostic system telling the user with a poor- looking
crop what the problem is—disease, some nutrient lacking, drought, etc.—is
a typical static system. An early ex- ample of such a system was used to
identify soybean insect problems and recommend treatment.
However, many agricultural problems may be used throughout a
growing season. Once the user enters data about the crop to assess the threat
of an early-season disease, for example, the user may want to do it a week
later to check on the same disease or on a question such as irrigation needs.
These subsequent uses can take advantage of the earlier data entered to give
decision support based on the history of the crop up to that point. These
programs are called dynamic decision sup- port systems, and they must keep
earlier data from one use to the next.