Study Of The Logistics Process Modeling And Optimization Methods

This paper is research on the modeling analysis and optimization computations of inventory control on logistic management. The logistics management and inventory control have become main roles of modern enterprises management. The positions and functions of the inventory control in the environment of supply chain are discussed first. The analysis on main conflict problems needed to be determined is presented. Based on the algebraic Petri nets the modeling research on vender managed inventory (VMI) is proceeded on. Considering the shortages of the original algorithms, the improvements are putout to speed up the search process and improve the precision of the search results. The optimizations of lots sizing problem with the objective function of inventory costs are processed by the improved optimal algorithms. By the analysis on the results of the optimization, the conditions of overtime work, out of stock, and security stock are achieved. The main novel points included are as following:The research on algebraic specifications of a logistics system is presented. By the analysis on the processes of vender managed inventory, the semantic specifications of vender managed inventory are achieved. The structure analysis of vender managed inventory is processed through the data dictionary, data flow diagram, and decision tables methods. Then the algebraic specifications of vender managed inventory are written.The modeling method with algebraic Petri nets is primarily used in inventory management problems. A procedure with the structure analysis method is induced to implement formal specifications of Petri nets from informal semantic scenario specifications. By using the data flow diagram, data dictionary, decision table of the vender managed inventory system and the transition oriented method, the algebraic Petri nets of vender managed inventory is written. The processes of the vender managed inventory are analysis by the algebraic Petri nets specifications.Two improve methods of particle swarm optimization algorithm are put forward. The disturbances are introduced to speed and positions to prevent premature phenomenon, convergence on local optimal values, and improve the search speed of the algorithm. Three neighborhood topologies are introduced to the algorithm to speed up each particle to convergence on the local optimal value for the complex multimodal functions. The simulations on several benchmark functions have proved the efficiency of the improved particle swarm optimization algorithms.Based on the original model of lots sizing problems, four models which is suitable for different circumstances are introduced by considering the out of stock costs and security stockcosts. They are overtime work costs included model, out of stock costs included model, securitystock costs included model, and all costs included model.Two models of the lots sizing problems which are considering the overtime work costs and out of stock costs have been processed to optimization search by the improved particle swarm optimization algorithm and improved genetic algorithm respectively. The data of the optimal costs, the production plan, and the inventory amounts are achieved from the search results. The strategies for constrains of stock balances and capacity balances are presented in the implementation processes of the algorithms. The criterions for the plan of overtime work and security stock are achieved from the analysis on the capacity demand in the search results.