An optimization model for vehicle routing problem in last-mile delivery

Due to rapid urbanization, timely delivery using vehicle routing is the most pressing issue for E-commerce logistics and distribution. In this study, we articulate multiple vehicle routing problems with a maximum capacity constraint and no time constraint. A brief literature review is conducted to identify the widely used optimization models for perishable goods delivery in the last mile. We have implemented and compared optimization algorithms such as Intra-Route Local Search, Inter-Route Local Search, and Tabu Search that provide a suboptimal solution to the greedy solution of this NP-hard problem. We compared the solution provided by these algorithms with the optimal solution that can be obtained in exponential time using processing time and precision. The results demonstrate that Tabu search outperforms other techniques for larger instance sizes, but for smaller instance sizes, Local search can produce results that are comparable to TABU search in a significantly shorter amount of time. In addition, the impact of instance size on the performance of the aforementioned algorithms is evaluated. The real-life evaluation is done to understand the use cases of this problem for an e-commerce company that will repeatedly face this challenge during warehouse storage management, last-mile delivery planning, and execution. A comprehensive performance comparison of various algorithms is presented to optimize last-mile delivery for future cases in order to reduce the overall carbon footprint and achieve profitability through the use of sustainable modes of distribution.

Automated summary

Due to rapid urbanization, timely delivery is the most urgent issue for E-commerce logistics and distribution. This study addresses multiple vehicle routing problems with a maximum capacity constraint and no time constraint. Various optimization algorithms are implemented and compared, and their solutions are compared with the optimal solution obtained through exponential time processing. Results show that Tabu search outperforms other techniques for larger instance sizes, but for smaller instance sizes, Local search can produce comparable results in a significantly shorter time. The impact of instance size on algorithm performance is evaluated, and a real-life evaluation is conducted to understand the use cases of this problem for an e-commerce company.