Enhancement of heuristic optimization algorithms based on evolutionary computational methods and applications

Abstract

This study aims to enhance the heuristic optimization algorithms, specifically the particle swarm optimization (PSO) and genetic algorithm (GA), to solve complex problems, namely the helicopter routing problem for inspecting high-voltage transmission lines of the Electricity Generating Authority of Thailand (EGAT) for maximum cost savings and identifying atmospheric pollution point sources. The first problem is the multi-depot vehicle routing problem (MDVRP), which involves additional special conditions such as avoiding backlit flights, fuel limitations, and helipad conditions. The study proposes an adaptive GA by enhancing the creation of the initial population, fitness function, and crossover operation, resulting in significant improvements in computational efficiency and generating outstanding results compared to the IBM CPLEX Optimizer. The enhanced GA is tested on 12 different problems and applied to the EGAT problem, providing effective results with low computational costs. The second problem seeks to improve the PSO algorithm by combining a multidimensional search with a line search and GA to detect air pollution point sources. The innovation, called HPSO, uses theoretically expected concentration from approximating the numerical solution of some governing PDE system to achieve better measurement results. HPSO outperforms PSO and GA in detecting air pollution point sources. In conclusion, The proposed adaptive GA and HPSO algorithms offer promising results with low computational costs, providing a valuable tool for solving similar problems.