Coordinating lot sizing and integrated production and distribution scheduling with batch delivery and holding cost

PurposeThe aim of this paper is to present a method for finding the optimum balance between sequence-dependent setup costs, holding costs, delivery costs and delay penalties in an integrated production–distribution system with lot sizing decisions.Design/methodology/approachTwo mixed integer linear programming models and an optimality property are proposed for the problem. Since the problem is NP-hard, a genetic algorithm reinforced with a heuristic is developed for solving the model in large-scale settings. The algorithm parameters are tuned using the Taguchi method.FindingsThe results obtained on randomly generated instances reveal a performance advantage for the proposed algorithm; it is shown that lot sizing can reduce the average cost of the supply chain up to 11.8%. Furthermore, the effects of different parameters and factors of the proposed model on supply chain costs are examined through a sensitivity analysis.Originality/valueAlthough integrated production and distribution scheduling in make-to-order industries has received a great deal of attention from researchers, most researchers in this area have treated each order as a job processed in an uninterrupted time interval, and no temporary holding costs are assumed. Even among the few studies where temporary holding costs are taken into consideration, none has examined the effect of splitting an order at the production stage (lot sizing) and the possibility of reducing costs through splitting. The present study is the first to take holding costs into consideration while incorporating lot sizing decisions in the operational production and distribution problem.

Integrated Production and Distribution Problem of Perishable Products with a Minimum Total Order Weighted Delivery Time

In this paper, an integrated production and distribution problem for perishable products is presented, which is an NP hard problem where a single machine, multi-customers, and homogenous vehicles with capacity constraints are considered. The objective is to minimize the total order weighted delivery time to measure the customer service level, by making two interacted decisions, production scheduling and vehicle routing, simultaneously. An integrated mathematical model is built, and the validity is measured by the linear programming software CPLEX by solving the small-size instances. An improved large neighborhood search algorithm is designed to address the problem. Firstly, a two-stage algorithm is constructed to generate the initial solution, which determines the order production sequence according to the given vehicle routing. Secondly, several removal/insertion heuristics are applied to enlarge the search space of neighbor solutions. Then, a local search algorithm is designed to improve the neighbor solutions, which further generates more chances to find the optimal solution. For comparison purposes, a genetic algorithm developed in a related problem is employed to solve this problem. The computational results show that the proposed improved large neighborhood search algorithm can provide higher quality solutions than the genetic algorithm.