Bio-Inspired Modifications of the PSO Algorithm

In PSO, a population of individuals, referred to as "particles", move through the search space, adjusting their positions and velocities based on both their individual experiences and the experiences of the swarm as a whole. This combination of individual exploration and swarm cooperation helps guide the particles towards optimal solutions. This algorithm has several advantages: it is easy to describe and implement, requires a relatively small number of function evaluations to converge, and boasts a fast rate of convergence. It has undergone numerous variations and improvements, including modifications to the update equations, incorporation of constraints, and hybridization with other optimization techniques. In this thesis, we will introduce bio-inspired modifications to the PSO algorithms, follow- ing the considerations given in Section 1.3. While Particle Swarm Optimization is highly effective and suitable for modeling swarms of animals, from a biological perspective, this algorithm needs to undergo slight changes, that will be presented and described in detail in Chapter 2 . The rest of the thesis is organized as follows. The formalization of a generic PSO will be presented in Chapter 1, including a description of the algorithm, and comments on the component ingredients and parameters. As previously introduced, Chapter 2 will be dedicated to presenting our proposed method. In Chapter 3 we will describe our two objective functions and numerical settings used in our simulations. Moreover, we will explore how different values of the model coefficients affect simulation outcomes, specifically the method’s ability to reach convergence for both functions. Finally, in Chapter 4, we will present a variation of our proposed method, with slight changes in some terms of the algorithm.