Loading...

Multi-depot Vehicle Routing Using Predictive Distributed Cooperative Method

Rahimi Baghbadorani, Reza | 2021

205 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54461 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Haeri, Mohammad
  7. Abstract:
  8. Due to conspicuous increases in shopping and goods transportation in this unbridled world, a significant part of companies’ final interest and final cost are related to their distributions of goods. Meanwhile, because of practical constraints of service hours and service distances, logistics companies usually build multiple depots to serve a great number of dispersed customers. Hence, multi-depot vehicle routing problems (MDVRP) win more and more engineers’ and companies’ attention to thrift their cost by using optimization methods. Considering NP-hardness of MDVRP and the lack of optimal solution derived by heuristic and meta-heuristic approaches in the large scale problems, in this project, first, we address the classic form of MDVRP and ponder its solution by breaking the problem into small ones and transferring them into several TSPs or CVRPs helping us to solve the problem providently. Also, a control parameter is introduced to trade-off between final cost and computation time. Then, regarding dynamic issues in the real world, such as traffic congestion, increasing or decreasing number of customers, and other uncertainties marring the solution achieved by classic MDVRP, we analyze dynamic MDVRP using a model predictive approach and its features, specially receding horizon one, where at least one part of the problem is changed by the passage of time. Using this method helps us to solve the problem in terms of uncertainties again bringing opportunities for vehicles to serve customers belonging to other vehicles previously, illustrating the co-operative feature, reduce the computational complexity and effect of uncertainties. Finally, the solutions of classic and dynamic MDVRP are investigated and comparisons are carried out between proposed methods in this project and other existent methods for solving classic MDVRP. Furthermore, the performance of dynamic MDVRP used the concept of predictive control is illustrated in three scenarios besides classic MDVRP. Experimental results show that the proposed algorithm is feasible and efficient to solve both classic MDVRP and dynamic one to the extent that computational time is decreased dramatically in classic MDVRP and the final cost is plunged about fifteen percent in dynamic MDVRP compare to a classic problem
  9. Keywords:
  10. Traffic Congestion ; Predictive Control ; Multi-Depots Location ; Capacitated Clustering ; Classic Multi-Depot Vehicle Routing Problems (MDVRP) ; Dynamic Multi-Depot Vehicle Routing Problems (MDVRP) ; Good Order

 Digital Object List

 Bookmark

No TOC