Research On Modeling Of Third-party Logistics Based On Petri Net

In recent years,Third-party Logistics has been serving more and more companies at home and abroad for its advantages such as multiple functions,low costs,and its high efficiency in management and control of logistics information.Based on the reported studies on Third-party Logistics,its modelling,scheduling and delivery are addressed in this thesis.The main results are given as follows.Firstly,an open-loop generalized stochastic Petri net model of a Third-party Logistics system is obtained and its system performance is analyzed.The closed-loop model is presented in the aim of solving deadlock and providing method of handling abnormal condition.Based on the closed-loop model,the two most critical processes that affected the operational efficiency of the system are identified as the receiving and delivery processes.The two processes are quantitatively analyzed to find out how they affect the system when they change in efficiency.Secondly,the problem of how to maximize the system efficiency by scheduling tasks and allocating resources is studied in the case of multi-tasking and multi-operation teams working simultaneously.Based on the obtained generalized stochastic Petri net models of parallel task groups,three scheduling strategies and two resource allocation strategies for group multi-task parallel operations without sharing resources and shared resources are studied using Python simulation.The results show that the shortest latency scheduling strategy is the most robust scheduling strategy,and the shortest latency resource allocation method performs best.The delivery plan based on genetic algorithms is finally addressed in this thesis.With the aim of overcoming the disadvantages of the classical genetic algorithm,the result of the evolution of the population is closer to the optimal value at a faster speed via improving the initial population forming method,the cross-mutation operation and a single individual optimization.