SOLVING DETERMINISTIC MULTI-OBJECTIVE ASSIGNMENT PROBLEMS USING ANT COLONY OPTIMIZATION ALGORITHM

Abstract

This study is proposing a metaheuristic algorithm known as Ant Colony Optimization Algorithm (ACOA) to solve deterministic Multi-Objective Assignment Problems (MOAPs). While single-objective assignment problems have only one objective to optimize, MOAPs have several conflicting objectives to optimize simultaneously, such as minimizing production time and maximizing the productivity. ACOA is well-suited to solve this type of problems to a certain degree of accuracy because of the adaptive search and the collaborative learning.  In this study, the ACOA is applied to seek a feasible near optimal solution to the deterministic assignment model, where in the implementation of the algorithm, the artificial ants construct assignments based on pheromone trails and heuristic desirability derived from the probabilistic estimates. The mechanism of pheromone evaporation and probabilistic path selection permit the algorithm to locate the near optimal solution while adhering to a set of constraints. The proposed algorithm is coded in Python programming language and the efficiency of the algorithm in solving MOAPs is compared with an existing prominent method known as Technique for an Order of Preference by Similarity to Ideal Solution (TOPSIS). The proposed Multi-Objective Ant Colony Optimization Algorithm (MOACOA) is proven to be capable of solving large scale MOAPs in less computational time compared to the TOPSIS method. The extension of this study can focus on solving stochastic MOAPs.