Article Preview
Top1. Introduction
A typical optimization problem of engineering is the economic load dispatch (ELD) problem for fossil fuel based electrical power generation system which meets the lion’s share of the power demand of industry and domestic segment even in this 21st century. Though the entire world is concerned about the environmental impact of burning fossil fuel and also fast depletion of the natural reserve of fossil fuel, feasibility of widespread use of hydel, solar and other renewable energy sources is not established globally. Till such time that use of these alternative power sources are made wider, cheaper and also easier technologically, the importance of fusel fuel will be predominant and its optimal usage and economy will continue to be the central point of research in power generation industry. Besides this basic reason for continued research on ELD problem, and allied CEED and DEED problems, it has been consistently targeted by the soft computing community owing to its typical combinatorial nature and unlimited possibilities for improvement of optimality using heuristic methods and their combinations. Every little improvement in fuel cost and wastage or loss minimization is important from the point of view of ever-increasing scarcity and cost of fossil fuel.
Many population-based soft computing methods are used to solve ELD problem; some of these methods are Genetic Algorithms (GA) (Holland, 1992; Swarup & Yamashiro, 2002; Chiang, 2009), Particle Swarm Optimization (PSO) (Yuan et al., 2009; Jeyakumar et al., 2006), Ant Colony Optimization (ACO) (Hou et al., 2002), Biogeography-based Optimization (BBO) (Simon, 2008), Gravitational Search Algorithm (GSA) (Rashedi, 2009; Hota & Sahu, 2015; Duman et al., 2010), Spiral Optimization algorithm (SOA) (Benasla et al., 2014), Exchange Market algorithm (EMA) (Ghorbani & Babaei, 2016), Rooted Tree Optimization (RTO) (Labbi et al., 2016), etc.
A recent trend in the solution of ELD problems is the application of population-based heuristic algorithms in hybrid mode, thereby creating the opportunity to exploit the strength of each algorithm and mitigate the individual weaknesses. Few such recently applied hybrid methods that have given promising result in simulated ELD problems of small to midscale systems were closely studied and their salient points and research significance are briefly discussed below.