Optimization of Dynamic Economic Load Dispatch Using Bat Algorithm

Abstract

Dynamic Economic Load Dispatch (DED) is a real time power system problem, which is an extension of conventional economic load dispatch. DED is considered to minimize the fuel cost of the generating station considering equality and non-equality constraints. In this dissertation work, to optimize the DED problem, BAT algorithm (BA) has been implemented. Bat algorithm, inspired by nature, imitates the echolocation property of micro-bats found in nature. Loudness and rate of pulse are variable factors through which global best solution is being reached. Parameter control, frequency tuning and automatic zooming are some of the advantages provided by bat algorithm. In this work for further improvement in the exploration and exploitation capabilities of BA, the characteristics like inertia weight and impact factors has been introduced in the BA. The inertia weight factor is implied for improving the convergence speed and impact factor is implied for protecting the best solution from being struck around the local minimum. Two test system has been considered for checking the effectiveness of these modifications over the standard algorithm. Test system-Ⅰ consists of the five thermal generating units and test system-Ⅱ consists of the ten thermal generating units taking into consideration their generation level constraint and valve-point loading effect for duration of 24 hours. Performance of modified bat algorithm is compared with standard bat algorithm and other state of art algorithms i.e. harmony search algorithm, evolutionary algorithm and sequential quadratic programming. Convergence characteristics of modified and standard bat algorithm is presented to consolidate the robustness of the algorithm.