GENETIC ALGORITHM-BASED MACHINE LEARNING CLASSIFIER SYSTEM MODEL FOR SHORT-TERM UNIT COMMITMENT PROBLEM

摘要

This article proposes a genetic algorithm-based machine learning classifier system (MLCS) model to solve the short-term unit commitment problem. The proposed model utilizes the machine learning-classifier system to determine the near optimal unit commitment schedule for a given study period. The MLCS model consists of three different layers of classifier systems running in parallel for low, medium, and high consumer loads. A simple genetic algorithm (SGA) with crossover, mutation, and advanced operators such as dominance and diploidy is used for each set of classifiers. In the proposed model, each chromosome in the condition part and the action part is encoded in the form of a position-dependent gene representing the load demand and the unit commitment schedule. The allele value of the chromosome in the action part gives the ON/OFF state of the units as a commitment decision at each time period. To demonstrate the effectiveness of the proposed model, the authors conduct extensive studies 10, 26, and 34 generating unit systems.