TABU SEARCH

D.Vanitha,
(2015209042)
Control & Instrumentation
Engineering.
 The word “Tabu”
• Comes from Tongan, a polynesia language to
indicate the things that cannot be touched
because they are sacred
• Webster dictionary  “a prohibition imposed
by socialcustom as a protective measure” or
of something “banned as constituting a risk.”
 Glover’s Definition
• The overall approach is to avoid entrainment
in cycles by forbidding or penalizing moves
which take the solution, in the next iteration,
to points in the solution space previously
visited
 Factors Affecting Tabu Search
• Size
• Variability
• Adaptability of the tabu memory
 Applications
• Integer programming problems
• Scheduling
• Routing
• Traveling salesman problem
• Minimum spanning tree problem
 Basics of Tabu Search
• A heuristic search method
• Tabu restricts some search of neighboring
solutions
• Aspiration allows exception of the tabu
• Accessible Solutions
The solutions that are not in the tabu list,
or in the tabu list but satisfy aspirationconditions.
 Tabu Search
• Tabu search is a local search strategy with a
flexible memory structure
• Adaptive memory
• Responsive exploration strategies
• Always move to the best available
neighbourhood solution point, even if it is
worse than the current solution point.
 Tabu Search
• Maintain a list of solution points that must be
avoided (not allowed) or a list of move
attributes that are not allowed. This is referred
to as the tabu list.
• Update this list based on some memory
structure (short-term memory).
• Aspiration Criteria
Allow for exceptions from the tabu list, if
such moves lead to promising solutions.
 Tabu Search
• Diversification and Intensification:
• Diversification: Search the unexplored area of
the solution space by
Increasing tabu tenure
Changing tabu restriction
• Intensification: Candidate locations are
identified the search is more focused to
produce local optimal solutions
 Tabu Search
• Long Term Memory:
Frequency based memory/Recency based
memory
• Adaptive Memory Programming (AMP)
 Basic Tabu Search Algorithm
• Step 1: Choose an initial solution i in S. Set i* = i and k=0.

• Step 2: Set k=k+1 and generate a subset V of solution in N(i,k) such that
either one of the Tabu conditions is violated or at least one of the
aspiration conditions holds.

• Step 3: Choose a best j in V and set i = j.

• Step 4: If f(i) < f(i*) then set i* = i.

• Step 5: Update Tabu and aspiration conditions.

• Step 6: If a stopping condition is met then stop. Else go to Step 2.

 Example
– Minimum spanning tree problem with constraints.
– Objective: Connects all nodes with minimum costs

Costs
B B
20 30 20 30

10 5 10 5
A C E A C E
25 25
15 40 15 40
D D

An optimal solution without

20+10+15+5=50 considering constraints

Constraints 1: Link AD can be included only if link DE also is included. (penalty:100)

Constraints 2: At most one of the three links – AD, CD, and AB – can be included.
(Penalty of 100 if selected two of the three, 200 if all three are selected.)
 Example
Iteration 1 Add Delete Cost
Cost=50+200 (constraint penalties) BE CE 75+200=275
B BE AC 70+200=270
20 30 BE AB 60+100=160
10 5 CD AD 60+100=160
A C E
CD AC 65+300=365
25 DE CE 85+100=185
Delete 15 40 Add
D DE AC 80+100=180
DE AD 75+0=75

New cost = 75 (iteration 2)

( local optimum)

Constraints 1: Link AD can be included only if link DE also is included. (penalty:100)

Constraints 2: At most one of the three links – AD, CD, and AB – can be included.
(Penalty of 100 if selected two of the three, 200 if all three are selected.)
 B
20 30

10 Add Delete Cost

A C E
BE CE 75+200=275
25 BE AC 70+200=270
15 40 BE AB 60+100=160
D
CD AD 60+100=160
20+10+15+30=75 75+200=275 CD AC 65+300=365
DE CE 85+100=185
B DE AC 80+100=180
20 30 DE AD 75+0=75

A C 5 E
25
15 40
D
20+30+5+15=70 70+200=270
B
20 30

10 5
A C E
25
20+10+5+40=75 75+0=75 40
D
 Tabu list: DE Example
Iteration 2 Cost=75
Delete
B Add Add Delete Cost
20 30 AD DE* Tabu move
AD CE 85+100=185
10 5
A C E AD AC 80+100=180
25 BE CE 100+0=100
15 40 Tabu
D BE AC 95+0=95
BE AB 85+0=85
CD DE* Tabu move
B CD CE 95+100=195
30

10 5 * A tabu move will be considered only if it would result in a

A C E
better solution than the best trial solution found previously
25 (Aspiration Condition)
15 40
D Iteration 3 new cost = 85 Escape local optimum

Constraints 1: Link AD can be included only if link DE also is included. (penalty:100)

Constraints 2: At most one of the three links – AD, CD, and AB – can be included.
(Penalty of 100 if selected two of the three, 200 if all three are selected.)
 Example
Tabu list: DE & BE Add Delete Cost
Iteration 3 Cost=85 AB BE* Tabu move
AB CE 100+0=100
AB AC 95+0=95
B Tabu
20 30 AD DE* 60+100=160
AD CE 95+0=95
10 5
A C E AD AC 90+0=90

25 Add CD DE* 70+0=70

15 40 Tabu CD CE 105+0=105
D
Delete * A tabu move will be considered only if it would result in a
better solution than the best trial solution found previously
(Aspiration Condition)
Iteration 4 new cost = 70 Override tabu status

Constraints 1: Link AD can be included only if link DE also is included. (penalty:100)

Constraints 2: At most one of the three links – AD, CD, and AB – can be included.
(Penalty of 100 if selected two of the three, 200 if all three are selected.)
 Example

B Optimal Solution
20 30
Cost = 70
10 5 Additional iterations only find
A C E
25 inferior solutions
15 40
D
THANK YOU....!!