Hybrid metaheuristics for the integrated and detailed scheduling of production and delivery operations in no-wait flow shop systems

Abstract

In today's complex business environment, the optimisation of operational processes is an essential factor for the survival of organisations. Many companies struggling to maintain competitiveness search for better operational methods and models to minimise costs and improve market service levels. In these companies, the coordination between manufacturing and logistics activities is critical. The integration and synchronisation of these activities are alternatives that companies require to improve their operational performance and, consequently, increase customer service levels. This article contributes to this approach by proposing and evaluating new heuristic methods to solve the problem of scheduling production and distribution operations in no-wait flow shop systems integrated with transportation systems which operate with different distribution strategies. Based on the assumption that the improvement in synchronisation and coordination of production and transportation operations minimises delivery times and increases service levels, we propose and evaluate heuristic methods for the problem. These heuristic methods implement scheduling algorithmic structures based on the iterated greedy (IG), iterated local search, simulated annealing (SA), and variable neighbourhood search metaheuristics, which jointly schedule manufacturing and transportation operations. In this context, the decision level is operational. The minimum total flow time is the main optimisation criterion. The problem studied herein is inspired on the real-life case of an electronic goods company. The analysis of the experimental results demonstrates the potential of these methods to solve the problem of different models of transportation. Specifically, the IGSA2 method obtained the best results, improving the quality of the initial solutions by up to 17% in small, medium, and large instances depending on the adopted initialisation heuristic. The best proposed algorithm IGSA2 was compared with the two state-of-the-art algorithms to obtain superior results. The study also verified a proposed hypothesis to improve the quality of the solutions generated using heuristic methods, which was proved to be true for most of the experiments.