A Multi -Objecti ve Opti mizati on Model to Support Freshly Cut Vegetable Processing Decisions

Abstract

This study presents a multi -objecti ve opti mizati on approach for decision making in fresh-cut vegetable processing, opti mizing processing ti mes and costs through the selecti on of alternati ve processes at various stages of the producti on. Despite the limited att enti on given to the fresh-cut vegetable industry, parti cularly in applying multi -objecti ve opti mizati on methods to support processing decisions, this study addresses the research need. The stages of freshcut vegetable processing, including peeling, cutti ng, washing, and packing, off er multi ple alternati ve methods with varying costs and processing ti mes. The problem is formulated as an integer bi-objecti ve combinatorial opti mizati on model aimed at opti mizing total processing ti me and cost. Two algorithms, the discrete non-dominati ng sorti ng geneti c algorithm- II (NSGA II) and the discrete non- dominated sorti ng parti cle swarm algorithm (NPSO), were applied to explore their complementary algorithmic performance. The local search behaviour of NSGAII was enhanced through several innovati ve local search operators including crossover, and mutati on operators, while various positi on and velocity update operators were used in NPSO. Both primary and secondary data were uti lised in esti mati ng the process parameters of each alternati ve processing methods. The results showed that NPSO exhibited more robust convergence, while NSGA-II produced a greater number of soluti ons in the Pareto front.