An iterative procedure for production-inventory-distribution routing problem

摘要

In this paper, the integrated Production, Inventory and Distribution Routing Problem (PIDRP) is modelled as audone-to-many distribution system, in which a single warehouse or production facility is responsible for restocking audgeographically dispersed customers whose demands are deterministic and time-varying. The demand can be satisfied from either inventory held at the customer sites or from daily production. A fleet of homogeeous capacitated vehicles for making the deliveries is also considered. Capacity constraints for the inventory are given for each customer and the demand must be fulfilled on time, without delay. The aim of PIDRP is to minimize the overall cost of coordinating the production, inventory and transportation over a finite planning horizon. We propose a MatHeuristic algorithm, an optimization algorithm made by the interpolation of metaheuristics and mathematical programming techniques, to solve the model. In this paper, we propose a two-phase solution approach to the problem. Phase I solves a mixed integer programming model which includes all the constraints in the original model except for the routing constraints. The model is solved by using Concert Technology of CPLEX 12.5 Optimizers with Microsoft Visual C++ 2010. In phase 2, we propose a variable neighborhood search procedure asudthe metaheuristics for solving the problem. Computational experiment is conducted to test the effectiveness of theudalgorithm.