Mathematical Modeling Of Transport Logistics

Problems
Oksana Pichugina
Mathematical Modelling and Artificial Intelligence Dept
National Aerospace University ”Kharkiv Aviation Institute”
Kharkiv, Ukraine
oksanapichugina1@gmail.com

Abstract—A real-world problem of transport logistics, which answers only one question – how to reduce the expected costs
generalities the generic problem in this area of research, is of fuel used to serve a given group of consumers. In fact, actual
formulated. Its mathematical model as a problem of combina- expenses will depend on the mileage, as well as on the degree
torial optimization in a combinatorial setting is formed. Issues
with a reformulation of the problem as Euclidean combinatorial to which the vehicle was loaded during the corresponding
optimization one are discussed as well as perspectives of its route. Another important item of delivery costs is the cost
solution by methods of nonlinear programming, discrete and of repair and maintenance of vehicles, which also depends on
continuous. the total mileage and the average load of vehicles.
Keywords—transport logistics, transportation problem, In this paper, we consider the practical problem of orga-
sales-man problem, combinatorial setting, Euclidean
combinatorial optimization nizing operational delivery by one vehicle, constructing its
model as a combinatorial optimization problem on a set of
I. INTRODUCTION permutations, as well as its mathematical model as a construct-
ing its models as a combinatorial optimization problem on a
Transportation Logistics, TL is a research field w i th its set of permutations and as a constrained quadratic Boolean
practical applications dealing with a search for optimal traffic optimization problem. In addition, a brief overview of the
flows [1]–[11]. necessary theoretical information on Euclidean combinatorial
Typically, the primary goal of a transport logistics problem optimization theory [14]–[16] is given.
(TL-problem, TLP) is to minimize the cost of delivery of a cer-
tain volume or weight [1]–[4], [6]–[11]. One way to solve this II. PROBLEM STATEMENT
problem is to solve the traveling salesman problem (a traveling Let a private entrepreneur A possesses one unit of freight
salesman problem, TSP) [7], [12], [13], which determines the transport of capacity V (m3 )/carrying capacity W (kg), de-
sequence of service for specific d e livery p o ints, w h ich h a s a liverers on a regular basis certain products within a city for
minimum bypass length. In mathematical modeling, various a company B, gets a fee depending on how much cargo he
TSP generalizations, such as the multiple salesman problem transported. A delivery plan is formed daily, depending on
(Multiple Traveling Salesman Problem, mTSP) [12], [13], can orders made by shops of the company B and transport involved
also be useful. on that day delivery, in addition to the vehicle of entrepreneur
Another well-known mathematical model of numerous A. It is required to find a p lan f or t he d ay-time d elivery by
transport logistics problems is the Transportation Problem entrepreneur A of company B from a central depot to shops
(TP). With its the help, one can perform long-term planning C1 ..., Cn , which orders of volumes/weights of v1 ..., vn (m3 )/
of cargo delivery carried out by a whole vehicle running from w1 ..., wn (kg) The goal is maximizing the expected profit of
a manufacturer or a depot (a storage, a warehouse) to delivery entrepreneur A from his delivering activity (further referred to
points (shops, stores, outlets) [6], [7], [10], [11]. as Problem 1).
Nowadays, in Ukraine, delivery management by successful Taking into account the way in which the entrepreneur’s
manufacturing companies usually is carried ”on the wheels”, services are paid, as well as the fact that, according to the
when products are manufactured in a wide assortment, the conditions, he has no choice whether to perform the delivering
storage stage is omitted, the product is loaded directly into a or not, and which trading points to serve, this problem can
vehicle and sent to their destinations. Thus, in this situation, be formulated as follows: it is necessary to deliver the given
application of the classical TP-model becomes problematic products the the prescribed destinations and, at the same time,
and requires considering modifications s u ch a s multi-product minimize the cost of delivery, which include fuel costs, repairs,
TP (Multi-index TP, MTP) [11] and other generalizations of depreciation and taxes. It is this goal that we choose as the
the TP-model. basis for formalizing this problem.
The application of the generic TLP-model in operational Suppose that, during T days of working for firm B, en-
logistics planning also causes certain difficulties, b ecause this trepreneur A has collected a data, including: mileage ΣD
model does not take into account the existing constraints and (in km), cost of fuel used ΣF (in units), maintenance and

978-1-7281-9083-9/20/$31.00 ©2020 IEEE

Authorized licensed use limited to: Auckland University of Technology. Downloaded on December 22,2020 at 01:31:21 UTC from IEEE Xplore. Restrictions apply.
repair cost ΣR (in units), depreciation and taxes ΣS (usd), the permutation from a set Jn = {1, ..., n} (further, set Pn ). Thus,
transported cargo volume ΣW (in tons) Also, there are known it is necessary to solve a problem of the form of (2), where
distances from the company’s central depot to its shops and F RC(π) : Pn → R1 provided that:
between these shops.
π ∈ Pn . (3)
Remark 1. From the available statistics, the following values
can be determined: Let
ΣD
• the average daily mileage D = (km);
T D = (dij )i,j∈Jn0 (4)
ΣF
• expected (average) daily expenses: a) total F =
T be a matrix of distances from the central depot C0 of firm
ΣR B to the shops Ci , i ∈ Jn and between these trading points
(usd); b) service and repair expenses R = (usd) c)
T themselves, where Jn0 = Jn ∪ {0}.
ΣS
depreciation and taxes cost S = (usd); Our model will be based on the combinatorial and algebraic
T
ΣW (Euclidean) statements of TSP [10], [12], so before moving on
• the average workload at the beginning of a day w 1 =
T to formalizing F RC in terms of Pn -elements and Euclidean
(in tons); average daily workload as the average between space vectors, we give some information from combinatorial
the average workload at the beginning of a day and zero optimization, including Euclidean combinatorial optimization
ΣW
at the end of the day w = w1 /2 = (t.); (ECO) [14]–[20].
2T
F ΣF
• average fuel cost per 1 km f =
D
=
ΣD
(usd); III. BACKGROUND
• average service and repair costs for transportation of 1 kg
R The current state of combinatorial optimization [21]–[24]
of cargo over a distance of 1 km r = (usd). allows us to say with confidence that for the effective solution
1000D · w
of combinatorial optimization problems, the most promising
An entrepreneur hired a decision-making specialist to de-
is the search for Euclidean statements of extreme combinato-
velop a daily strategy for delivering the cargo based on these
rial problems, which include modeling feasible combinatorial
statistics and daily information about the shops served by the
regions as subregions of Euclidean space and their mapping in
company’s own fleet of vehicles and the volume of their orders.
this space, which is also called ”embedding” into Euclidean
The entrepreneur sees the following resources for saving his
space. The availability of Euclidean formulations is closely
expenses and, respectively, increasing its profit. On the one
connected with the combinatorial nature of feasible regions,
hand, it is desirable to reduce the overall mid-day mileage of a
as well as with the possibility of an adequate representation
vehicle resulting in reducing fuel costs. On the other hand, the
of the problems parameters by a single numerical parameter
costs of service and repair will decrease if the vehicle moves
or by finite sequences.
as little as possible overloaded, in other words, if the average
In Euclidean statements of combinatorial optimization prob-
load of the vehicle during the working day decreases compared
lems, the discreteness of the feasible region allows sufficient
to what was observed by now. Depreciation expenses and the
flexibility of the recording forms of the mathematical model
amount of taxes per day are considered constant. Thus, the
and an option of transition from one of its version to another,
objective function:
respectively, and from one group of applied optimization
T C = F C + RC + DC → min, (1) methods to another, makes it possible to choose more effec-
tive ones in some cases. Among equivalent formulations, a
where TC is transportation cost, FC is fuel cost, RC is repair special role is played by the so-called ”continuous” or ”alge-
cost, DC is depreciation cost and taxes. braic” statements in which the feasible region is represented
Given that DC is a constant, namely, DC = S, it can be analytically, respectively, and it becomes possible to apply
seen that the problem (1) is equivalent to the following: classical nonlinear programming methods to its solution. In
F RC = F C + RC → min, (2) turn, those of them that make it possible to consider the
objective function and functional constraints as convex are of
where F RC is the cost of fuel, service and repairs. particular importance. When the immersion is carried out, it
We construct a mathematical model of the problem under becomes possible to consider continuous relaxation, such as
the assumption that the capacity and carrying capacity of polyhedral, obtained by replacing the condition of membership
vehicle A is sufficient to serve all the shops C1 ..., Cn and of a discrete region by the condition of membership of a
deliver their orders for one trip, i.e., polyhedron, which is its convex hull. Accordingly, there is
n
X n
X a need to study the properties of combinatorial polyhedra,
vi ≤ V, wi ≤ W. their analytical description and face structure. In turn, the latter
i=1 i=1 involves the study of the properties of a polyhedron graph.
This means that we aim to find the order of the transporta- Consider a generic combinatorial optimization problem in
tion service for the shops. Clearly, it can be represented by a a combinatorial setting (further, COP), where minimization is

Authorized licensed use limited to: Auckland University of Technology. Downloaded on December 22,2020 at 01:31:21 UTC from IEEE Xplore. Restrictions apply.
chosen as the optimization direction, ∃ m: I(Π) = Jm , that where dij is the weight of the arc (i, j) ∈ E (further referred
is, its form - to as TSP.M2).
If G is a complete graph, TSP takes the form (8) (9) and
φ(π) → min, (5)
turns into an unconstrained TSP. Besides, a real TSP can
π ∈ Π0 , (6) always be formulated in this form, because for this it is enough
0
Π = {π ∈ Π : ϕi (π) ≤ 0, i ∈ Jm }. to set a sufficiently large length for the missing arcs, ensures
their absence in the optimal route. This fact makes meta-
Its Euclidean combinatorial setting (further, ECOP) is its heuristic approaches very popular when solving TSP in the
equivalent reformulation in the form of: form of TSP.M2 [26], [27].
f (x) → min, A TSP version is also known, where G is complete, and its
fi (x) ∈ E 0 ⊂ Rn , vertices are points on a plane or in three-dimensional space,
that is, V = {xi }i∈Jn ⊂ Rν , ν = 2.3. As a result, the objective
E 0 = {x ∈ E : fi (x) ≤ 0, i ∈ Jm0 }, function (8) takes the form:
where E, E 0 are images of sets Π, Π0 . n−1
P
Given the complexity of combinatorial structures to be φD (π) = kxπ1 − xπn k + kxπi − xπi+1 k → min, (10)
i=1
optimized, the need to take into account the relationship
between their components, Euclidean settings are known only and the setting (9), (10) is called Euclidean TSP (further
for the insignificant class COPs [25]. Most classic COPs are referred to as ETSP) [28].
formulated as COPs of the form (5), If we introduce the notation yi = xπi i ∈ Jn , ETSP can be
represented as:
π ∈ Π,
n−1
X
that is, in the form of the unconstraint COP problem (UCOP), φD (π) = ky1 − yn k + kyi − yi+1 k → min,
where the condition Π0 = Π is satisfied. As Π combinatorial i=1
sets often serve, such as the set of permutations Pn , the y = (yi )i∈Jn ∈ Pn (V),
n
{0, 1} = {0, 1}n or their subsets [21]–
Q
Boolean set Bn =
i=1 where Pn (V) is the set of permutations of vectors with
[24], [26], [29], [30]. In the first case, COP belongs to the class V (further referred to as TSP.M3) [17]–[19]. According to
of optimization problems on permutations (permutation-based [18], [19], s-set corresponding to Pn (V) is a basic C-set of
optimization problems, PBPs) [31], [32], in the second - to permutations of ν-vectors [17]–[20] to the solution of various
the class of Boolean problems (Boolean problems, BPs) [22], problems in which, it is possible to apply the properties of this
[23], [30]. set and functions defined on it, found in the these and other
sources [34]–[39].
A. TSP: various mathematical models
For TSP, there exists the well-known ”algebraic” form of
TSP belongs to the quite narrow class of combinatorial op- the model (Miller-Tucker-Zemlin formulation of TSP), [25]:
timization problems that allow formulations like COP, ECOP,
n
PBPs and BPs. Accordingly, TSP allows, when solving it, to f (x) =
P
dij xij → min; (11)
use and combine the properties of the corresponding classes i,j=1,i6=j
of combinatorial sets with the properties of Euclidean space. n
P
xij = 1, j ∈ Jn ; (12)
So, the goal of the generic (asymmetric) TSP in the graphs i=1,i6=j
is to find the shortest Hamiltonian cycle in a weighted directed Pn
graph xij = 1, i ∈ Jn ; (13)
j=1,i6=j
G = {V, E}, |V| = n, |E| = m (7) ui − uj + nxij ≤ n − 1, i, j ∈ Jn \{1}, i 6= j; (14)
with a weight matrix D = (dij )i,j ∈ R n×n
formulated as ui ∈ Jn−1 , i ∈ Jn \{1}; (15)
COP: xij ∈ {0, 1}, i, j ∈ Jn , (16)
φD (π) → min statement as ECOP (further referred to as TSP (ECO1)),
0 representing it as a discrete programming problem on the set
subject to (6) (further referred to as TSP.M1), where Π is the
set of Hamiltonian cycles in G, φD (π) is the length of the of vectors (x, u) = (vec(X), (ui )i∈Jn )T of dimension n2 + n,
most of whose coordinates are Boolean, the rest are integer
Hamiltonian cycle π. TSP belongs to the class of PBPs and
(hereinafter, X = (xij )i,j∈Jn , vec(X) is the vectorization
allows another COP statement of the form:
operator of the matrix X).
n−1
P
φD (π) = dπn π1 + dπi πi+1 → min, (8)
i=1 IV. MATHEMATICAL MODELING OF PROBLEM 1
π = (π1 , ..., πn ) ∈ Pn , (9) Suppose that the costs RC linearly depend on a vehicle A
(πn , π1 ), (πi , πi+1 ) ∈ E, i ∈ Jn−1 , mileage, the same as F C (further referred to as Problem 2).

Authorized licensed use limited to: Auckland University of Technology. Downloaded on December 22,2020 at 01:31:21 UTC from IEEE Xplore. Restrictions apply.
Then optimization of the function (2) is equivalent to opti- In the notation introduced and taking into account (27), the
mizing the Hamiltonian cycle length from the depot of the value RC is represented as follows:
company B through the shops and back to the depot, provided n−1
X
that the constraint (9) holds. Problem 2 itself is equivalent RC = r Wi dπi πi+1 → min, (28)
to the TSP with the distance matrix (4), where the starting i=0
element of the cycle is A0 .
where r is the expected cost of a repair of vehicle A after
We adapt the mathematical models TSP.M2 and TSP transporting 1 kg of cargo per 1 km.
(ECO1) to Problem 2 and get: Combining the problem (3), (28) with the combinatorial
• a combinatorial formulation of Problem 2 (further re- model of Problem 2, we obtain the combinatorial model of
ferred to as P2.M2): Problem 1 (further referred to as P2.CO):
n−1 n−1
X
P
φD (π) = dπn π0 + dπi πi+1 → min, (17) F RC = sφD (π) + r Wi dπi πi+1 → min, (29)
i=0
i=0
subject to constraint (3); where s is the average cost of a fuel unit per 1 km mileage
• a Euclidean statement of Problem 2 (further referred to for a vehicle A.
as P2(ECO)): Denote πn+1 = π0 and take into account (17), (26). As a
n
P result, formula (29) becomes:
f (x) = dij xij → min; (18) n n−1 i
i,j=0 X X X
F RC = s dπi πi+1 + r (W − wπj )dπi πi+1 =
x00 = 1; (19)
n i=0 i=0 j=1
xij = 1, j ∈ Jn0 ;
P
(20) n
X n−1
X
i=0 =s dπi πi+1 + rW dπi πi+1 −
n
Jn0 ; i=0 i=0
P
xij = 1, i ∈ (21) (30)
j=0 n−1
XX i n
X
ui − uj + nxij ≤ n, i ∈ Jn0 ; (22) r wπj dπi πi+1 = (s + rW) dπi πi+1 −
i=0 j=1 i=0
ui ∈ Jn , i, j ∈ Jn0 ; (23) n−1 i
xij ∈ {0, 1}, i, j ∈ Jn0 ,
XX
(24) − rWdπn πn+1 − r wπj dπi πi+1 → min .
i=0 j=1
Conditions (18)-(24) generalize constraints (11)-(16) with
replacing Jn → Jn0 , C1 ..., Cn → C0 , C1 ..., Cn , and fixing C0 As a result, P1.CO has the form of (3), (30).
as the starting and last point of the route. Let us analyze ways of constructing the algebraic formula-
tion of Problem 1 (further referred to as P1.ECOP). To do this,
Let us move to modeling of the following subproblems,
one needs to formalize the objective function involved in (30)
which can be singled out from Problem 1:
in Cartesian coordinates. Unlike Problem 2, in the Euclidean
RC → min (25) statement, Problem 1 will not be linear anymore due to the
presence of the last term in (30).
subject to the constraint (3).
CONCLUSION
Sections of a route are enumerated according to its starting
point and are denoted Di , j ∈ Jn0 , respectively. Let W be the At first g lance, t he t ransportation l ogistic p roblem formu-
total weight of the cargo transported by vehicle A, and Wi be lated in this paper, pursues the interests of only one person –
the cargo transferred through a section Di , i ∈ Jn0 . Then the entrepreneur A. However, this is not quite true. Having
solved this problem, simultaneously, the activity of other
n
X people, settlements and involved trading points are improved.
W= wi Indeed, reducing the mileage of the vehicle implies that the
i=1
delivery will conduct in a shorter time, the products will go
. to the shop faster and will, expectedly, be sold faster. Thus,
In the D0 section, the entire load is transferred to D1 – wπ1 the financial s ituation o f t he s hops w ill a lso i mprove. O n the
and so on. After all, at the Dn section, the vehicle mileages other hand, a state of roads will deteriorate to a lesser extent.
a single mileage. Thus, in general, we have: Respectively, they will not need to be repaired so often and
i
X not in such a volume, for example, instead of capital, it will
Wi = W − wπj , i ∈ Jn0 , (26) be possible to limit patching. In the end, people who receive
j=1 the desired product faster will be satisfied, a nd t heir morale
will improve. Moreover, the volume of exhaust gases will be
in particular, reduced, respectively, the state of human health will improve.
Wn = 0. (27) Thus, it makes sense at the level of city authority to encourage

Authorized licensed use limited to: Auckland University of Technology. Downloaded on December 22,2020 at 01:31:21 UTC from IEEE Xplore. Restrictions apply.
in every possible way such entrepreneurs who systematically [22] B. Korte and J. Vygen, Combinatorial Optimization: Theory and Algo-
and reasonably approach the issue of careful use of both their rithms, 5th ed. Heidelberg ; New York: Springer, 2012.
[23] C. H. Papadimitriou and K. Steiglitz, Combinatorial Optimization:
own and urban property, and reduce the cost of rent and certain Algorithms and Complexity, Unabridged edition. Dover Publications,
taxes to them. In turn, city authorities and relevant city services 2013.
may be required about road users and optimize traffic in order [24] A. Schrijver, Combinatorial Optimization: Polyhedra and Efficiency.
Springer Science & Business Media, 2002.
to reduce downtime in traffic j ams, a t s tops a t t raffic lights [25] V. K. Leontev, Discrete optimization, Computational Mathematics and
and in areas with limited bandwidth. Mathematical Physics, vol. 47, no. 2, pp. 328–340, Feb. 2007, doi:
10.1134/S0965542507020157.
[26] E. Gimadi and M. Khachay, Extremal Problems on Sets of Permutations.
REFERENCES Ekaterinburg: UMC UPI, 2016. [in Russian]
[27] I. V. Sergienko, L. F. Hulianytskyi, and S. I. Sirenko, Classification of
[1] J. H. Bookbinder, Handbook of global logistics : transportation in applied methods of combinatorial optimization, Cybern Syst Anal, vol.
international supply chains. New York, NY : Springer, 2013. 45, no. 5, p. 732, Sep. 2009, doi: 10.1007/s10559-009-9134-0.
[2] U. Clausen, M. Ten Hompel, and M. Klumpp, Efficiency and logistics. [28] J. E. Yukich, Probability Theory of Classical Euclidean Optimization
Berlin ; New York : Springer, 2013. Problems. Berlin, Heidelberg: Springer-Verlag, 1998.
[3] J. R. Daduna, Vehicle scheduling, in Encyclopedia of Optimization, C. [29] S. Butenko, P. M. Pardalos, and V. Shylo, Eds., Optimization Methods
A. Floudas and P. M. Pardalos, Eds. Springer US, 2008, pp. 4027–4032. and Applications : In Honor of Ivan V. Sergienkos 80th Birthday. Cham:
[4] B. Fahimnia, M. G. H. Bell, D. A. Hensher, and J. Sarkis, Green Springer International Publishing, 2017.
logistics and transportation : a sustainable supply chain perspective. [30] P. M. Pardalos, D.-Z. Du, and R. L. Graham, Eds., Handbook of
Cham : Springer, 2015. Combinatorial Optimization, 2nd ed. New York: Springer, 2013.
[5] H. Gleibner and K. Mller, Case studies in logistics. Wiesbaden: Gabler [31] M. Mehdi, Parallel Hybrid Optimization Methods for permutation based
Verlag, 2011. problems, Phd[Thesis], Lille : Universite des Sciences et Technologie
[6] J. Gonzalez-Feliu, F. Semet, and J.-L. Routhier, Sustainable urban de Lille, 2011.
logistics : concepts, methods and information systems. Berlin : Springer, [32] J. Gmys, Heterogeneous cluster computing for many-task exact opti-
2013. mization - Application to permutation problems, Phd[Thesis], Mons,
[7] L. Gouveia, A. Paias, and S. Vo, Models for a traveling purchaser Lille : Universite de Mons (UMONS); Universite de Lille, 2017.
problem with additional side-constraints, Computers & Operations [33] R. A. Brualdi, Combinatorial matrix classes. Cambridge : Cambridge
Research, vol. 38, no. 2, p. 550–558, 2011. University Press, 2006.
[8] T. Gudehus and H. Kotzab, Comprehensive logistics. Heidelberg : [34] O. Pichugina and N. Muravyova, A Spherical Cutting-plane Method
Springer, 2009. With Applications In Multimedia Flow Management, in Proceedings of
[9] E. Taniguchi and R. G. Thompson, City logistics : mapping the future. the 1st International Workshop on Digital Content & Smart Multimedia
Boca Raton : CRC Press, 2015. (DCSMart 2019), Lviv, Ukraine, Dec. 2019, pp. 82–93.
[10] C. Vogiatzis and P. M. Pardalos, Combinatorial Optimization in Trans- [35] S. Yakovlev, O. Kartashov, O. Pichugina, and K. Korobchynskyi, Ge-
portation and Logistics Networks, in Handbook of Combinatorial Op- netic Algorithms for Solving Combinatorial Mass Balancing Problem,
timization, P. M. Pardalos, D.-Z. Du, and R. L. Graham, Eds. Springer in 2019 IEEE 2nd Ukraine Conference on Electrical and Computer
New York, 2013, pp. 673–722. Engineering (UKRCON 2019) Conference Proceedings, Lviv, 2019, pp.
[11] M. Queyranne and F. Spieksma, Multi-index transportation problems, 1061–1064.
in Encyclopedia of Optimization, C. A. Floudas and P. M. Pardalos, [36] B. Farzad, O. Pichugina, and L. Koliechkina, Multi - Layer Community
Eds. Springer US, 2008, pp. 2413–2419. Detection, in 2018 International Conference on Control, Artificial
[12] M. Choi and S. Lee, The multiple traveling purchaser problem for Intelligence, Robotics Optimization (ICCAIRO), May 2018, pp. 133–
maximizing systems reliability with budget constraints, Expert Systems 140, doi: 10.1109/ICCAIRO.2018.00030.
With Applications, vol. 38, no. 8, pp. 9848–9853, Aug. 2011. [37] S. Yakovlev, O. Pichugina, and O. Yarovaya, On Optimization Prob-
[13] R. E. Burkard and E. ela, Linear Assignment Problems and Extensions, lems on the Polyhedral-Spherical Configurations with their Properties,
in Handbook of Combinatorial Optimization, D.-Z. Du and P. M. in 2018 IEEE First International Conference on System Analysis
Pardalos, Eds. Springer US, 1999, pp. 75–149. Intelligent Computing (SAIC), Kyiv, Oct. 2018, pp. 94–100, doi:
[14] Y. G. Stoyan and O. Yemets, Theory and methods of Euclidean 10.1109/SAIC.2018.8516801.
combinatorial optimization. Kiev: ISSE, 1993. [in Ukrainian] [38] O. Pichugina, Placement problems in chip design: Modeling and
[15] Y. G. Stoyan and S. V. Yakovlev, Mathematical models and optimiza- optimization, in 2017 4th International Scientific-Practical Conference
tion methods in Geometric Design. Kiev: Naukova Dumka, 1986. [in Problems of Infocommunications. Science and Technology (PIC S T),
Russian] Oct. 2017, pp. 465–473, doi: 10.1109/INFOCOMMST.2017.8246440.
[16] Y. G. Stoyan and S. V. Yakovlev, Theory and Methods of Euclidian [39] O. Pichugina and O. Kartashov, Signed Permutation Polytope Pack-
Combinatorial Optimization: Current Status and Prospects, Cybern Syst ing in VLSI Design, in 2019 IEEE 15th International Conference
Anal, vol. 56, no. 3, pp. 366–379, May 2020, doi: 10.1007/s10559-020- on the Experience of Designing and Application of CAD Systems
00253-6. (CADSM) Conference Proceedings, Lviv, 2019, pp. 4/50-4/55, doi:
[17] O. Pichugina and S. Yakovlev, Quadratic Optimization Models and 10.1109/CADSM.2019.8779353.
Convex Extensions on Permutation Matrix Set, in XIV International
Scientific and Technical Conference Computer Science and Information
Technologies, Lviv, Ukraine, Sep. 2019.
[18] O. Pichugina and S. Yakovlev, Euclidean Combinatorial Configurations:
Continuous Representations and Convex Extensions, in Lecture Notes
in Computational Intelligence and Decision Making, Zalizniy Port,
Ukraine, Jul. 2019, pp. 65–80, doi: 10.1007/978-3-030-26474-1 5.
[19] O. Pichugina and S. Yakovlev, Euclidean Combinatorial Configurations:
Typology and Applications, in 2019 IEEE 2nd Ukraine Conference
on Electrical and Computer Engineering (UKRCON 2019) Conference
Proceedings, Lviv, 2019, pp. 1065–1070.
[20] Y. G. Stoyan and I. V. Grebennik, Combinatorial species for the
enumeration of combinatorial configurations with special properties,
Dopov. Nats. Akad. Nauk Ukr. Mat. Prirodozn. Tekh. Nauki, no. 7,
pp. 28–32, 2010. [in Russian]
[21] W. J. Cook, W. H. Cunningham, W. R. Pulleyblank, and A. Schrijver,
Combinatorial optimization. John Wiley & Sons, Inc., New York, 1998.

Authorized licensed use limited to: Auckland University of Technology. Downloaded on December 22,2020 at 01:31:21 UTC from IEEE Xplore. Restrictions apply.