A Study On Joint Replenishment And Delivery Modelswith Stochastic Demand Using Differential Evolution Algorithm

Reasonable inventory management can help an enterprise achieve a higher servicelevel with relatively lower cost, thus enhance enterprise’s competitiveness greatly. It isquite common for enterprises to jointly replenish items from multi-suppliers ormulti-places, which can evidently reduce the annual ordering times and decrease thetransportation cost. Therefore, joint replenishment policy can be an effective way for costsavings. However, Joint Replenishment Problem (JRP) has been proven to be an NP-hardproblem and the key is to find efficient and effective algorithma. This thesis discussedseveral JRPs under stochastic demand and the algorithms to solve the proposed models.Firstly, a differential evolution algorithm (DE) is adopted to overcome the shortagesof current approaches. A hybrid DE algorithm (HDE) which integrates the advantages ofGA is designed and its performance is tested by four classic testing functions. Secondly,HDE-based approach for the JRP under stochastic demand is proposed. Results ofconstrative examples indicate that HDE is effective and robust. Sensitivity analysis forthree parameters has been presented to obtain some managerial inspiration. Thirdly, amodel integrates multi-buyer joint replenishment and distribution (JRD) with stochasticdemand and the diagram of HDE-based algorithm is given. Numerical example indicatesthat HDE is better than DE, GA and the best heuristic in terms of convergence rate andoptimal solution. Moreover, HDE is easy to be implemented with excellent accuracy androbustness. Finally, a multi-objective model with the minimum total cost and stock outobjectives based on JRD is presented. Two solutions have been discussed by linearprogramming and multi-objective evolution algorithm (MOEA). Results of numericalexample show the simplicity and efficiency of MOEA.