Optimal Solution of the Dynamic Economic Dispatch by Improved Teaching-Learning-Based Artificial Bee Colony Algorithm

Abstract

Dynamic economic dispatch is one of the most handled problem in modern power system operations. It aims to optimize the output power from thermal generating units over a specified time period to minimize the total fuel cost, while satisfying the several constraints such as generation limits, ramp rate limits, and power balance. In addition to these constraints, the prohibited operating zones and the valve-point loading effect are included the DED problem. In this case, the complexity, nonlinearity, and non-convexity of the DED problem are increases. Therefore, in order to solve the DED problem, a powerful meta-heuristic search (MHS) algorithm are proposed. In this study, an improved teaching-learning-based artificial bee colony (TLABC) algorithm, where the fitness-distance balance based TLABC (FDB-TLABC) and natural-survivor method based TLABC (NSM-TLABC) algorithms were hybridized. To prove the performance of the proposed algorithm, it was applied to solve the DED problem and benchmark problem suites. In the simulation study carried out on benchmark problems, the results of the proposed algorithm and five MHS algorithms were evaluated statistically. According to Friedman test results, the proposed algorithm ranked first with 2.2836 values among them. On the other hand, the proposed algorithm and its rival algorithms were applied to solve the two DED cases. The results of them show that the proposed algorithm achieved superior performance to find the best objective values for both case studies.