Optimization of Horticultural Food Commodity Distribution Routes using Genetic Algorithm with Crossover Partially Match

Abstract

Horticultural food commodities have unique characteristics: perishable and not durable. In addition, the distribution of these commodities is also uncertain along highly complex routes. We experimented to determine the route of potato distribution in Central Java, Indonesia. We use metaheuristic methods to find solutions to these problems. This method solves the problem using algorithms for optimization. This study aimed to determine the most optimal route using a genetic algorithm (GA) with partially matched crossover (PMX). The GA stages in this study include (1) Population Initialization; we use the Random Generator Algorithm, (2) Selection; we use the Roulette Wheel Selection method, (3) Crossover; we use the PMX method and (4) Mutation. We used ten chromosomes with seven genes each. We used ten chromosomes with seven genes each. The results of this study indicate that the best route is obtained in the second generation, namely through node1 - node7 - node2 - node6 - node5 - node3 - node4 - node1 with an optimal distance of 159 km. For further research, consider the spatial conditions for a better solution.