See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228949723

Using genetic algorithms and simulation as decision support in marketing

strategies and long-term production planning

Conference Paper · September 2009

CITATIONS READS

8 256

2 authors:

Florin Stoica Laura Florentina Stoica

Lucian Blaga University of Sibiu Lucian Blaga University of Sibiu
57 PUBLICATIONS 117 CITATIONS 33 PUBLICATIONS 44 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Florin Stoica on 05 June 2014.

The user has requested enhancement of the downloaded file.

 Proceedings of the 9th WSEAS International Conference on SIMULATION, MODELLING AND OPTIMIZATION

Using genetic algorithms and simulation as decision support in

marketing strategies and long-term production planning
FLORIN STOICA LAURA FLORENTINA CACOVEAN
Computer Science Department Computer Science Department
Faculty of Sciences Faculty of Sciences
University “Lucian Blaga” Sibiu University “Lucian Blaga” Sibiu
Str. Dr. Ion Ratiu 5-7, 550012, Sibiu Str. Dr. Ion Ratiu 5-7, 550012, Sibiu
ROMANIA ROMANIA
florin.stoica@ulbsibiu.ro laura.cacovean@ulbsibiu.ro

Abstract: - This paper represents an approach of using simulation models and genetic algorithms for generating an
aggregative production plan to maximize the total profit of the firm. The described methodology provides a tool which
assists the company management in decision which of the suitable solutions will become the production plan. The
entire system is composed by a business information system - a database, a simulation model and a genetic algorithm.
The purpose of the integrated system is to help operative management personnel to take decisions with respect to long-
term production planning and marketing strategies.

Key-Words: - Production planning, Genetic algorithms, Co-mutation

1 Introduction is calculated with a fitness function. After being

Most real life optimization and scheduling problems are evaluated with a fitness function, each chromosome in
too complex to be solved completely. The complexity of population receives its fitness value.
real life problems often exceeds the ability of classic The optimization is based on a genetic algorithm
methods. In such cases decision-makers prepare and which uses a new co-mutation operator called LR-Mijn,
execute a set of scenarios on the simulation model and capable of operating on a set of adjacent bits in one
hope that at least one scenario will be good enough to be single step.
used as a production plan. In present, there is a major interest in design of
A long time goal for scheduling optimization research powerful mutation operators, in order to solve practical
has been to find an approach that will lead to qualitative problems which can not be efficiently resolved using
solutions in a relatively short computational time. The standard genetic operators. These new operators are
development of decision-making methodologies is called co-mutation operators. In [7] was presented a co-
currently headed in the direction of simulation and mutation operator called Mijn, capable of operating on a
search algorithm integration. This leads to a new set of adjacent bits in one single step. In [12] we
approach, which successfully joins simulation and introduced and studied a new co-mutation operator
optimization. The proposed approach supports man- which we denoted by LR-Mijn and we proved that it
machine interaction in operational planning. offers superior performances than Mijn operator.
A group of widely known meta-heuristic search The paper is organized as follows. In Section 2 we
algorithms are genetic algorithms (GA). Evolutionary make a brief presentation of the architecture of our
algorithms are a very effective tool that enables solving decision-support system, called GA-SIM. We also
complicated practical optimization problems. An present the basic idea of evolutionary method adopted
important characteristic of evolutionary algorithms is for optimization process. In section 3 is introduced the
their simplicity and versatility. Their main drawback is co-mutation operator LR-Mijn. The evolutionary
a long calculation time, which, however, is not a serious algorithm based on the LR-Mijn operator, used within the
disadvantage nowadays with advanced computer decision-support system is presented in section 4.
technology and does not limit their use for searching for Section 5 contains the description of the simulation
almost optimal solutions, if are not real-time restrictions model and its integration in the GA-SIM system.
[2]. Conclusions and further directions of study can be found
With computer imitation of simplified and idealized in section 6.
evolution, an individual solution-chromosome represents
a possible solution to our problem. Chromosome fitness

ISSN: 1790-2769 435 ISBN: 978-960-474-113-7

 Proceedings of the 9th WSEAS International Conference on SIMULATION, MODELLING AND OPTIMIZATION

2 The architecture of the decision-support and according to defined criteria, the decision-maker is
motivated to search for results which will have the most
system advantageous influence on the whole production process.
Companies need to be flexible to compete for the market
The result, which is selected after simulation on the
share and adapt to market demands by offering
visual simulation model, becomes the proposed
competitive prices and quality. This calls for a wide
production plan
assortment of products or product types, small
production costs, etc. Simulation is a strong interactive
tool that helps decision-makers improves the efficiency Fitness function
of enterprise actions. The ability of simulation to show a
real process on the computer with the consideration of Simulation model

kth generation
uncertainty is a big advantage when analyzing system Business
Data
behavior in complex situations [15]. Information
System
Our system is composed by a business information
system - a database, a simulation model and a genetic New individual
evolution (co-
algorithm, and is called in the following the GA-SIM mutation)
system.
A simulation model will be used for fitness function
computation of genetic algorithm results, as well as for
visual representation of qualitative evaluation of a Result of fitness

fitness value
Individual
chosen production plan following genetic algorithm eval. (profit)
optimization.
The value of the fitness function is computed by the
simulation model, with uses the respective chromosome
Selection
as input data. The value returned from the simulation Last
model is used to evaluate that chromosome, which generation
encodes a possible production plan.
By applying the genetic operations on the members of Optimal
a population of decisional rules at the moment t, it production plan
results a new population of rules that shall be used at the
moment t+1. The population from the initial moment, t = Fig. 1 The architecture of the decision-support system
0 is generated randomly and the genetic operations are
applied iteratively until the moment T (at which the Every chromosome codes o possible production plan.
condition for stopping the algorithm is accomplished). The quality of a chromosome is represented by the
The above iterative process may be interpreted amount of profit which result from the simulation model
economically as it follows. The evolution process has as if it receives as input data the production plan coded in
objective the finding of the successful individuals. The that chromosome.
binary strings of these individuals (chromosomes) with It is noticed that, through the modification of the
high fitness values (a high profit) will be the basis for simulation model, it is obtained a more varied general
building a new generation (population). The strings with frame which leads to more diversified suggestions
lower fitness values, which represent decisions to concerning the decisions that refer to the quantities of
produce with a low profit, find few successors (or none) company products that shall be offered on the market.
in the following generation.
The purpose of the integrated system is to aid 3 The LR-Mijn operator
operative management personnel in production planning In this section we define the co-mutation operator called
and marketing strategies (incorporated in simulation LR-Mijn. Our LR-Mijn operator finds the longest
model). The main advantage of the presented system is sequence of σp elements, situated in the left or in the
to enhance man-machine interaction in production right of the position p. If the longest sequence is in the
planning, since the computer is able to produce several left of p, the LR-Mijn behaves as Mijn, otherwise the LR-
acceptable schedules using the given data and a set of Mijn will operate on the set of bits starting from p and
criteria. The user then selects the most suitable schedule going to the right.
and modifies it, if necessary.
Let us consider a generic alphabet A = {a1, a2, …, as}
After completion of the optimization process, the
most suitable production plans are simulated on the composed by s ≥2 different symbols. The set of all
visual model of the system. Using chosen parameters sequences of length l over the alphabet A will be

ISSN: 1790-2769 436 ISBN: 978-960-474-113-7

 Proceedings of the 9th WSEAS International Conference on SIMULATION, MODELLING AND OPTIMIZATION

denoted with ∑ = Al. where k ≠ i (randomly chosen).

In the following we shall denote with σ a generic (iii) If p ≠ right and p = l – left – 1,
string, and σ = σl-1…σ0 ∈ ∑= Al, where σq ∈ A ∀ q ∈ right , i
σ = σ ( p, i ) σ(p-right-1,n)... σ(0) and
{0, …, l-1}. Through σ(q,i) we denote that on position q left , i
within the sequence σ there is the symbol ai of the right , i
alphabet A. LR-Mijn(σ) = σ ( p, k ) σ(p-right-1,n)... σ(0), where k ≠ i
left , k
z
σ p, j denotes the presence of z symbols aj within the (randomly chosen).
sequence σ, starting from the position p and going left right , i
right , n (iv) If p = right and p = l – left –1, σ = σ ( p, i )
and σ ( p, i ) specify the presence of symbol ai on left , i
left , m
LR-Mijn(σ)=
position p within the sequence σ, between right symbols
an on the right and left symbols am on the left. We ⎧ right ,i

right ,i ⎪σ left = σ ( p , k ) , for left > right , where k ≠ i

suppose that σ = σ(l-1)... σ(p+left+1,m) σ ( p, i ) σ(p- ⎪ left , k

left ,i ⎪ right , k
⎨σrightt = σ ( p, k ) , for left < right , where k ≠ i
right-1,n)... σ(0). ⎪ left ,i
The Mijn operator is the mutation operator defined in ⎪σright or σrleft for left = right , with probability 0.5
[7]: ⎪
Mijn : σ ∈ ∑, p ∈ {0, … , l-1} → σ’ ∈ ∑’ ⊂ ∑, where p ⎩
is randomly chosen As an example, let us consider the binary case, the
Mijn string σ = 11110000 and the randomly chosen
(i) σ = σl-1…σp+nσp+n-1,i σ np,-1j σp-1…σ0 ⎯⎯⎯→
application point p = 2. In this case, σ2 = 0, so we have to
σ’ = σl-1…σp+nσp+n-1,j σ np,-i1 σp-1…σ0, for n < l – p + 1 find the longest sequence of 0 within string σ, starting
and from position p. This sequence goes to the right, and
Mijn because we have reached the end of the string, and no
(ii) σ = σ lp,- pj σp-1…σ0 ⎯⎯⎯→ σ’ = σ lp,- pk σp-1…σ0, occurrence of 1 has been met, the new string obtained
for n = l – p + 1, with ak ≠ aj randomly chosen in A. after the application of LR-Mijn is 11110111.
In [12], we introduced and study the properties of LR- The commutation operator LR-Mijn allows long
Mijn co-mutation operator jumps, thus the search can reach very far points from
where the search currently is. We proved in [12] that the
Definition 3.1 Formally, the LR-Mijn operator is defined LR-Mijn operator performs more long jumps than Mijn,
as follows: which leads to a better convergence of an evolutionary
(i) If p ≠ right and p ≠ l – left – 1, algorithm based on the LR-Mijn in comparison with an
algorithm based on the Mijn operator.
LR-Mijn(σ)= In the following, we will consider that A is the binary
⎧ right ,i alphabet, A = {0, 1}.
⎪σ left = σ (l − 1)...σ ( p + left + 1, i ) σ ( p, m)
⎪ left , m
4 The evolutionary algorithm based on
⎪ σ ( p − right − 1, n)...σ (0) for left > right
⎪ LR-Mijn operator
⎪ right , n
The basic scheme for our algorithm, called in the
⎨σrightt = σ (l − 1)...σ ( p + left + 1, m) σ ( p, n)
⎪ left ,i following LR-MEA, is described as follows:
⎪ σ ( p − right − 1, i )...σ (0) for left < right
⎪ Procedure LR-MEA
⎪σright or σrleft for left = right , with probability 0.5 begin
⎪⎩
t=0
(ii) If p = right and p ≠ l – left – 1, Initialize randomly population P(t) with P elements;
right ,i Evaluate P (t) by using fitness function;
σ = σ(l-1)... σ(p+left+1,m) σ ( p, i ) and while not Terminated
left ,i
for j = 1 to P-1 do
right , k
- select randomly one element among the
LR-Mijn(σ) = σ(l-1)... σ(p+left+1,m) σ ( p, k ) ,
left , i best T% from P(t);

ISSN: 1790-2769 437 ISBN: 978-960-474-113-7

 Proceedings of the 9th WSEAS International Conference on SIMULATION, MODELLING AND OPTIMIZATION

- mutate it using LR-Mijn; each member (chromosome) of the population.

- evaluate the obtained offspring; The probability to select a certain chromosome
- insert it into P’(t). (possible production plan) in the next generation is
end for related with its performance in simulated conditions (the
profit provided by the simulation model). That profit
Choose the best element from P(t) and determines the fitness value of the respective possible
solution of our optimization problem.
Insert it into P’(t)
P(t+1) = P’(t) Chromosome
t=t+1 (Possible production plan)
end while
end
Fitness function
Evaluation
5 The simulation model Simulation
module
Model
The simulation model is implemented as an Excel (Java)
(Excel app)
application. It is based on real data provided by the
Business Information System but also on forecasted data
(e.g. sales quantities and values for the following
months). The model take account of many dates and Profit
variables: sales (values and quantities), raw material
costs, discounts, packaging costs, direct & indirect labor JExcelAPI
costs, energy cost, depreciation from the rate of
exchange, warehousing and logistic costs, transport
costs, advertising & promotions, etc. The main Fig. 3 Implementation of the fitness function through the
workbook contains few sheets, a VBA module, and simulation model
results are synthesized in the following pivot table:
Because the GA-SIM system is implemented in Java
as main language, was necessary a bridge between Java
code and the simulation model, implemented in Excel.
Our choice for this purpose was JExcelAPI
(http://jexcelapi.sourceforge.net/), a mature, open source
java API enabling developers to read, write, and modify
Excel spreadsheets dynamically [14].

6 Conclusions and further directions of

study
The purpose of the GA-SIM integrated system is to aid
operative management personnel in production planning
and marketing strategies (incorporated in simulation
model). The main advantage of the presented system is
to enhance man-machine interaction in production
planning, since the computer is able to produce several
acceptable schedules using the given data and a set of
criteria.
The GA-SIM system is fully implemented, and
currently is under evaluation in a big company from
Sibiu, Romania, which was interested in its acquisition.
Fig. 2 A pivot table from the simulation model As a further direction of study we want to compare
the results obtained by using different genetic operators
The role of the simulation model in the genetic and to evaluate real codifications of variables, instead of
algorithm is presented in the Figure 3. current binary one.
In fact, the simulation model represents the
implementation of the fitness function of the genetic
algorithm, needed in the procedure LR-MEA to evaluate

ISSN: 1790-2769 438 ISBN: 978-960-474-113-7

 Proceedings of the 9th WSEAS International Conference on SIMULATION, MODELLING AND OPTIMIZATION

References: [12] Stoica F., Simian D., Simian C., A new co-
mutation genetic operator, Proceedings of the 9th
[1]Jaber A. Q, Hidehiko Y., F., Ramli R., Machine WSEAS International Conference on Evolutionary
Learning in Production Systems Design Using Computing, Sofia, Bulgaria, May 2008, pp. 76-81.
Genetic Algorithms, International Journal of [13] Vapnik V., The Nature of Statistical Learning
Computational Intelligence, No. 4, 2008, pp. 72-79. Theory, Springer Verlag, 1995.
[2]Kofjač D., Kljajić M., Application of Genetic [14] Java Excel API, http://jexcelapi.sourceforge.net
Algorithms and Visual Simulation in a Real-Case [15] Kljajić M., Bernik I., Škraba A., Leskovar R.,
Production Optimization, WSEAS TRANSACTIONS Integral simulation approach to decision assessment
on SYSTEMS and CONTROL, Issue 12, Volume 3, in enterprises, Shaping future with simulation:
December 2008, pp. 992-1001. proceedings of the 4th International Eurosim 2001
[3]Radhakrishnan P., Prasad V. M., Gopalan M.R., Congress, Delft University of Technology, 2001.
Optimizing Inventory Using Genetic Algorithm for
Efficient Supply Chain Management, Journal of
Computer Science 5 (3), 2009, pp. 233 - 241.
[4]Lo Chih-Yao, Advance of Dynamic Production-
Inventory Strategy for Multiple Policies Using
Genetic Algorithm, Information Technology Journal
7 (4), 2008, pp. 647-653.
[5]De Falco I., An introduction to Evolutionary
Algorithms and their application to the Aerofoil
Design Problem – Part I: the Algorithms, von
Karman Lecture Series on Fluid Dynamics,
Bruxelles, April 1997
[6]De Falco I., Del Balio R, Della Cioppa A.,
Tarantino E., A Comparative Analysis of
Evolutionary Algorithms for Function Optimisation,
Research Institute on Parallel Information Systems,
National Research Council of Italy, 1998
[7]De Falco I, A. Iazzetta, A. Della Cioppa, Tarantino
E., The Effectiveness of Co-mutation in Evolutionary
Algorithms: the Mijn operator, Research Institute on
Parallel Information Systems, National Research
Council of Italy, 2000
[8]De Falco I., Iazzetta A., Della Cioppa A., Tarantino
E., Mijn Mutation Operator for Aerofoil Design
Optimisation, Research Institute on Parallel
Information Systems, National Research Council of
Italy, 2001
[9]Chen J., Using Genetic Algorithms to Solve a
Production-Inventory Model, International Journal
of Business and Management, Vol. 2, No. 2, 2007,
pp. 38-41.
[10] Krishnakumar K., Goldberg D., Control system
optimization using genetic algorithm, Journal of
Guidance, Control, and Dynamics, no. 15(3), 1992,
pp. 735-740.
[11] Zhu X., Huang Y., Doyle J., Genetic algorithms
and simulated annealing for robustness analysis,
Proceedings of the American Control Conference,
Albuquerque, New Mexico, 1997, pp. 3756-3760.

ISSN: 1790-2769 439 ISBN: 978-960-474-113-7

View publication stats