The value of integrated planning for production, inventory, and routing decisions: A systematic review and meta-analysis

This paper presents a comparison of sequential and integrated planning for the production routing problem, in which production, inventory, and routing decisions must be made. The aim is to estimate the expected value of treating the problems as a whole, rather than making decisions sequentially. In particular, the following research questions are posed: What is the expected cost reduction when combining production, inventory, and routing in a single modeling framework, compared to solving the problems individually in a sequence? Under which circumstances is it most beneficial to tackle an integrated problem? In other words, the goal is to establish whether the solutions obtained by the integration are clearly better than approximate solutions obtained by a more simplified process, and if so, under which circumstances this difference is the most pronounced. To answer these research questions, a systematic review was performed, resulting in a set of 20 relevant articles that were analyzed in depth. For the first research question, computational results from 15 articles were obtained and analyzed through a meta-analysis. The analysis estimated an expected cost savings provided by integration of 11.08%, with a 95% confidence interval of [6.58%, 15.58%]. For the second research question, individual results obtained via sensitivity analyses in 20 relevant articles were summarized qualitatively, enabling insights into how the potential savings by integration is influenced by parameters such as the degrees of freedom, the cost, and the capacity.

Automated summary

This paper presents a comparison between sequential and integrated planning for the production routing problem, and explores the advantages and influencing factors of integrated problem-solving. Through a systematic review and analysis of 20 relevant articles, a qualitative summary of the cost savings rate and influencing factors of the integrated problem is obtained.