Interactive gravitational search algorithm and pattern search algorithms for practical dynamic economic dispatch

Abstract

Summary Dynamic economic dispatch strategy consists in determining the output of generation of each unit with respect to predicted load demand over a period of time satisfying the practical unit constraints. In dynamic operation of units, ramp rates (ramp down limit and ramp up limit) are the important constraints that affect considerably the life of the rotor of the generating unit. Recently, a new category of metaheuristic optimization methods has been introduced and attracted many researchers and expert engineers due to their capacity and efficiency to find the near global optimal solution. In this work, a new optimization technique known as gravitational search algorithm (GSA) is proposed and adapted in coordination with pattern search algorithm (PS) to solving the dynamic economic dispatch considering valve point loading effects, ramp rates of generating units and power losses. The proposed hybrid method named communicated GSA–PS allows balancing between exploitation and exploration capability, which makes agents to react more by changing experiences with local search mechanism. The proposed approach has been examined and applied to three practical power systems, 40 generating units to solving the static economic dispatch, 5 generating units, and 10 generating units considering ramp rates limits, valve point effect, and transmission power losses. From the different case studies, it is observed that the results compared with the other recent techniques demonstrate the particularity of the proposed approach and show clearly its effectiveness to solve practical dynamic ED problem. Copyright © 2014 John Wiley & Sons, Ltd.

Ant lion optimizer versus particle swarm and artificial immune system for economical and eco‐friendly power system operation

“…Recently, several evolutionary programs are investigated for solving ELD problem. The literature documents the most relevant . The evolutionary algorithms in general require limited knowledge of the problem.…”

“…The most promising meta‐heuristic approaches for ELD are fast converging evolutionary programming, ant colony, population variant differential evolution, genetic algorithm, Tabu search, differential harmony, bacterial foraging, artificial bee colony, firefly algorithm, simulated annealing, semi‐definite programming, gravitational search algorithm, and hybrid univariate marginal distribution algorithm . It could be concluded based on literature that some evolutionary approaches suffer from serious drawbacks as premature convergence, time consuming, and initial solution sensitivity. Moreover, they differ widely according to the implementation complexity, the computation speed, and the storage requirements …”

“…It could be concluded based on literature that some evolutionary approaches suffer from serious drawbacks as premature convergence, time consuming, and initial solution sensitivity. Moreover, they differ widely according to the implementation complexity, the computation speed, and the storage requirements …”

“…To remedy the deficiency of a single evolutionary technique and enhance the performance, the merging between two meta‐heuristic methods is advised. The synergy improves the search capabilities and overcomes the limitations of each method.…”

“…The approach of combining two evolutionary techniques provides better solution for ELD than that of single optimization method. However, the hybrid scenario is more complicated and requires more computation requirements than the elementary method …”