Harnessing the Power of Population-Based Hybrid Algorithms in Electrical Engineering

Abstract

Optimization problems pop up in a variety of fields, including civil, mechanical, and electrical engineering, as well as
telecommunications, chemical and biochemical engineering, computer science, manufacturing systems, economics,
agricultural sciences, physical sciences, networking systems, pattern recognition, security systems, and business
management, among others. Nowadays, the latest optimization techniques are being embraced across many areas of
human activity, especially where complex decisions need to be made that can be modeled mathematically. Essentially,
optimization is all about decision-making, where the goal is to enhance one or more objectives based on specific
circumstances. Most real-world problems tend to be non-linear, and they typically require a global optimal solution.
That's why many users prefer to develop optimization techniques that can pinpoint the global optimal solution instead
of just settling for a local one in non-linear scenarios. SOMA, a fresh addition to the family of population-based
heuristics, is gaining traction due to its broad applicability. However, it does face challenges, particularly when it
comes to the curse of dimensionality. Since SOMA doesn't generate new solutions during its search but merely shifts
existing positions, it can struggle with slow convergence as dimensions increase. This is largely due to its limited
diversity mechanism. To tackle this issue, SOMA can be combined with features from other methods, leading to the
creation of new algorithms known as variants of the original algorithm.