Network Optimization using Multi-Agent Genetic Algorithm

This paper present an innovative technique based on multi-agent genetic algorithm for optimization of a network. We unify agent system with genetic algorithm and applied to solve multi-objective problem optimization. In this algorithm an agent illustrate a candidate results to the optimization problem. Agent lives in the grid environment and it possesses own local space called the neighborhood. In the neighborhood, an agent can compete and collaborate with other agents, to attain the purpose of gene exchanged and evolved. Agent also possesses some cognition of the surroundings and can pursue itself while expands, with the aim to adapt itself to the surroundings better and increases its viability. A new multi-agent genetic algorithm is proposed named as MAGA-NOP, in which we implement crossover operator based on neighborhood to get useable information from its neighbor and by doing this we avoid it from random recombination. We used priority based encoding mechanism to encode chromosome strings. Several networks are used to test the algorithm performance; the experimental results revealed that MAGA-NOP has a progressive performance than other algorithms.