Loading...

Vehicle Routing Problems with Mobile and Stochastic Customer Locations (VRP-MSCL)

Shafaei, Abolfazl | 2024

0 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 57168 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Akbari Jokar, Mohammad Reza; Rafiee, Majid; Hemmati, Ahmad
  7. Abstract:
  8. In classical vehicle routing problems, a fixed and specific location is considered for each customer, which in many cases does not reflect reality. Many customers change their locations when distribution is carried out, and in some cases, it is challenging to predetermine their locations. This thesis proposes and develops the Vehicle Routing Problem with Mobile and Stochastic Customer Locations (VRP-MSCL), a subproblem of last-mile delivery. The concept of mobile and stochastic customer refers to a type of customer who is constantly moving, lacks a fixed location, and cannot be identified in advance; however, forecasting customers’ locations in advance is possible. This concept introduces a novel idea in the customer domain, which we have termed an “Addressless Customer.” We proposed a general model for this type of vehicle routing problem and explored approaches to address it. As a case study, we investigate spare parts distribution among roadside assistance cars in a company. Roadside assistance cars are constantly on the move and lack fixed locations, thus perfectly aligning with the VRP-MSCL model. Since each roadside assistance car’s service area is known, historical data facilitates estimating the probability of their presence in their respective service areas. Therefore, the spare parts distribution among roadside assistance cars can be modeled and solved using the proposed approach for solving VRP-MSCL problems. Employing a Mixed Integer Programming (MIP) model for vehicle routing problems with mobile and stochastic customer locations by using stochastic programming we attempt to determine vehicle tours among customers. The model was solved for the case study using CPLEX software and compared with a case with fixed locations and deterministic parameters. The proposed approach and its results, validated by managers and experts from the company, demonstrate its ability to enable the development of new customer services and potentially enhance satisfaction among roadside assistance car owners. The numerical results obtained from implementing this approach suggest a significant reduction (approximately 20%) in distribution costs compared to the current method. Furthermore, as vehicle routing problems are classified as NP-hard and require substantial time for exact solutions for medium to large-sized problems, a heuristic method is proposed specifically for these vehicle routing problems. This method has been applied to sample problems that could be solved by exact methods within a reasonable timeframe. The proposed method reached solutions approximately 65% faster on average than the exact method and reached optimal solutions in over 80% of cases
  9. Keywords:
  10. Vehicle Routing Problem ; Spare Parts ; Stochastic Programming ; Addressless Customer ; Spare Parts Distribution ; Mobile and Stochastic Customer Locations (MSCL) ; Roadside Assistance

 Digital Object List

 Bookmark

No TOC