REALIZING ARTIFICIAL IMMUNE SYSTEM IN A MULTI AGENT SIMULATION ENVIRONMENT AND AN APPLICATION TO TRAVELLING SALESMEN PROBLEM

BAYKASOĞLU A., Saltabas A., TAŞAN A. S., SUBULAN K.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.27, sa.4, ss.901-909, 2012 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 27 Sayı: 4
  • Basım Tarihi: 2012
  • Dergi Adı: JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.901-909
  • Anahtar Kelimeler: Agent based modeling and simulation, artificial immune system, clonal selection algorithm, travelling salesmen problem
  • Dokuz Eylül Üniversitesi Adresli: Evet

Özet

Although many meta-heuristic algorithms were developed for solving combinatorial optimization problems, very few of them were realized in an agent based environment. Especially the algorithms which model dynamics of Artificial Immune Systems (AIS) are population based approaches with adaptability characteristics, therefore AIS can be better realized in an agent based modeling environment. For this purpose first time in the literature a clonal selection algorithm which is an AIS based algorithm is modeled in a multi-agent environment for solving the travelling salesmen problem which is a combinatorial optimization problem. In order to observe the behavior of the algorithm, simulation experiments are carried out on several test problems. Netlogo software is utilized for developing agent based models and simulation tests. Moreover, receptor change process and crossover mechanisms are integrated into the proposed model in order to improve the performance of the classical clonal selection algorithm. It is shown that there is a high potential to obtain good solution by making use of agent oriented approaches which more realistically model the natural phenomenon.