Genetic Algorithms Based Inventory Simulation Optimization For Complex Logistics

Complex Logistics (CL) is a network logistic system of supply chain composed ofmany entities like manufactures, products suppliers and retailers. There exist someuncertain factors such as customer demands, product supplies, and payment cycle, whichcould make the system a multi-layer system mixed with both continuous and discreteevents. Because of the complexity of CL, it can not obtain the best order-up-to level ofeach entity in advance, and then the inventory management would be affected. The firsttask of inventory management is to reduce the cost.The existed integrated simulation software CLCSim has represented the whole theoryand methodology for modeling and simulation of CL, including structure of the model,establishment of the simulation system and accomplishment of the software. CLCSimcould solve the problem of the uncertainty of the CL system. It also can identify the bestorder-up-to level of each entity and decrease the inventory cost. However, there are alsoproblems like too many simulation results and too long simulation time. In this paper, onemethod applying genetic algorithm, which is to obtain the order-up-to level of CL, isintroduced. The aim of the solution is to reduce the inventory cost. On the basis of thesimulation optimization platform CLCSim, combined with the intelligence of geneticalgorithm, the new method could get the best order-up-to level through setting differentcontrol parameters of genetic algorithm. Meanwhile, after many groups of tests, theinfluence of different population size and different crossover rate is discussed. This willplay positive roles in the process of choosing proper order-up-to level, as well as providecritical and valuable reference to decrease the inventory cost of CL.