Heuristics with novel approaches for cyclical multiple parallel machine scheduling in sugarcane unloading systems

Abstract

This paper focuses on a computational tool for scheduling sugarcane vehicles for dump tippler machines operating at a sugar mill. This problem was defined as scheduling M parallel capacitated machines with a cyclic sequence where machine restriction, and sequencing independent setup time are included with the objective to minimise makespan. To solve the problem, mathematical programing was developed to solve small-sized problems, while realistic-sized problems were solved by an effective metaheuristic called variable neighbourhood strategy adaptive search (VaNSAS). According to our literature review, it is the first time that VaNSAS has ever been proposed to solve the problem. The procedure, formulae and components of VaNSAS were first created based on the idea of increasing the search performance of existing heuristics. Additionally, K-variable move heuristics have been also first proposed. The VaNSAS was developed using the traditional Differential Evolution (DE) with heuristics embedded in it to obtain near optimal conditions for solving realistic-sized problems. The numerical results showed that the VaNSAS outperformed all other proposed methods, since it could often find new optimal solutions during the simulation, while the local search based heuristics were often trapped at some local optima and the DE lacked search intensification.