Population Based Equilibrium in Hybrid SA/PSO for Combinatorial Optimization: Hybrid SA/PSO for Combinatorial Optimization

2. Background

Metaheuristics have been developed over the years as a means in which to tackle computationally difficult problems. One of the most well studied metaheuristics is SA, which was originally applied for solving the TSP in the work of (Černý, 1985). One of the core properties that makes SA so appealing as a metaheuristic is the ability for it to accept worse solutions that would have otherwise been discarded as they lead to poorer short-term solutions (Franzin & Stützle, 2019). Alternatively, SA is regarded as relying too heavily on initial parameters (Trelea, 2003; van den Bergh & Engelbrecht, 2006); thereby requiring significant domain-level and application-level knowledge of the problem before initializing. Researchers have stayed on top of this problem by introducing hybrid strategies that complement the shortcomings of SA alone. These strategies will be discussed in the following section.