Evaluating Chaotic Map Integration Strategies for MetaheuristicBased Team Formation Optimisation

Abstract

This paper proposes the Chaotic Roulette Wheel Zebra Optimization Algorithm (CRWZOA), which is a hybrid variant that has been designed for optimizing the healthcare team formation problem. In this algorithm, different types of chaotic maps are considered for initializing the population, such as Logistic, Tent, Sine, and Singer. Then, the population size for each type of chaos map is determined using the adaptive roulette-wheel selection approach based on the inverse average cost of the maps during the early search process. After that, the selected initialized population will be used to apply the standard foraging and defense phases in the ZOA framework. This algorithm was implemented on a healthcare team formation dataset containing 1000 people and 30 skills. Cost, team size, and execution time were considered as evaluation metrics. The results show that CRWZOA outperforms other ZOA variants and provides the lowest cost, smallest team size with zero variance (?² = 0) and lowest execution time in comparison to other variants of the ZOA algorithm. The results have confirmed that combining chaotic maps with roulette wheel selection is effective in solving complex problems like team formation in complex domains like healthcare.