A method to eliminate solution trapping in applying progressive optimality principle to short-term hydrothermal scheduling

Abstract

This paper presents a method to eliminate solution "trapping" for the P.O.P. (Progressive Optimality Principle) based short-term hydrothermal scheduling algorithm. In a P.O.P. based scheduling algorithm, the "trapping" phenomenon can severely hinder solution optimality, and thus limits its applicability. The "trapping" phenomenon is caused by the "weak" links which break the optimization path into isolated segments, and thus destroy the path continuity required by the P.O.P. based decomposition approach. In order to ensure path continuity, the proposed method identifies and bypasses the potential "weak" l i s , such that global optimality can be achieved progressively by successively solving a sequence of two-hour sub-problem. In this paper, examples are used to illustrate the "trapping" phenomenon and the proposed method.

Hydrothermal systems generation scheduling using cultural algorithm

“…Therefore, many methods have been developed to solve this problem over the past decades. The major methods include variational calculus (Grake & algorithm (Turgeon 1981;Lee 1989), Lagrangian relaxation method (Tufegdzic 1996;Salam & Mohamed 1998) and modern heuristics algorithms such as artificial neural networks (Naresh & Sharma 1999), evolutionary algorithm [17 -20] (Chen & Chang 1996;Yang & Yang 1996;Orero & Irving 1998;Werner & Verstege 1999), chaotic optimization (Yuan & Yuan 2002), ant colony (Huang 2001), Tabu search (Bai & Shahidehpour 1996) and simulated annealing (Wong & Wong 1994). But these methods have one or another drawback such as dimensionality difficulties, large memory requirement or an inability to handle nonlinear characteristics, premature phenomena and trapping into local optimum, taking too much computation time.…”