Abstract

The Unit Commitment (UC) problem refers to the process of optimal power generating unitsstart-up and shut-downschedule determination, subject to forecast load demand over a short-term planning horizon (24h). The objective of a generation scheduling optimization problem is to minimize total operating costs, while meeting a large set of system operating constraints. This problem includes two basic decisions--unit commitment decision and economic dispatch decision. The characteristics of this problem are high dimension, non-convex, discrete, nonlinear, and multi-constraints. Many optimization methods have been proposed to solve the UC problem. These methods included Priority List (PL) methods, Dynamic Programming (DP) methods,Lagrangian Relaxation (LR) methods. More recently, meta-heuristic methods have been tested and used, such as Genetic Algorithms (GA),Tabu Search (TS) and Simulated Annealing (SA), along with expert system and neural networks. Genetic Algorithms (GA) have become increasingly popular in recent years in science and engineering disciplines.Some works have been published covering the solution of the UCP and EDP using GA.A single technique like Fuzzy Logic itself is not sure to produce good results, but it can besupplemented with other method like Genetic Algorithm (GA).A new approach is proposed, a hybrid algorithm to solve the UCP and EDP.