A new intelligent multi-agent system for management of ordering policies in a fuzzy supply chain

In this paper a new intelligent multi-agent system is proposed for finding the best ordering policy. The best ordering policy is the policy which minimizes the total cost of the supply chain that is the sum of all echelons´ costs over all periods. The best ordering policy is obtained by a new window-base genetic algorithm. One limitation of the previous presented GA-based algorithms is the constraint of the fixed ordering rule for each member through the time. To resolve this problem a new concept -window- is introduced that is a parameter of the model. Application of the window basis enables the agents to have different ordering policies through the time. Another limitation of the previous research is the weak management of uncertainty. In this research, supply chain´s main parameters such as demand value, ordering amount, lead time and costs are all modeled by fuzzy numbers. The results show that the proposed multiagent system has a lower cost in comparison with similar research in the literature.