Study On Ant Colony Optimization Algorithm And It's Application To Optimal Generating Unit Commitment Of Power System

Nowadays, with the marketable restructure of power industry, it is more and more important to maintain the operation of electric utility economically and efficiently. The Unit Commitment (UC) optimization is the key problem of the optimal short-term generation scheduling just because it brings significant annual financial savings to power utilities. The unit commitment (UC) optimization is typical nonlinear optimization problems which rely on many factors, such as nonlinear problem, discrimination, stochastic effect, uncertainty and so on. When dealing with this kind of problem, the traditional mathematical methods have its own limitations. But the optimization method by simulated evolution is uniquely suited to solve these high nonlinearity problems, discrimination problems and uncertainty problems. In this paper, Ant Colony Optimization presented recently was applied to Unit Commitment (UC) optimization. Researches on Ant Colony Optimization and its applications in Generator Unit Optimal Commitment are presented.Ant Colony Optimization is a novel optimization algorithm of simulated evolutionary algorithms. It is improved to solve the prematurely and stagnation problems of traditional Ant Colony Optimization in the paper. First, mean value method is used to update pheromone to avoid premature convergence which leads to non-globally optimal solution. Second, Pheromone is updated by Best Track to speed the rate of convergence which will help to avoid locally optimal solution.In this paper, we focused on the Thermal Power Generator unit commitment (UC) optimization. A mathematical model is presented and consuming minimization of generator is chosen for its objective function, with the power balance, generation capacity limits, maximum allowable start and stop times limits and time limits taken into consideration.The unit commitment (UC) optimization is constrained minimization problem. Different methods are used to deal with these restrictions in this paper. We changed the unit commitment (UC) optimization into unconstrained minimization problems which solved by Ant Colony Optimization. In the end, examples are used to show the application of Ant Colony Optimization in Generator Unit Optimal Commitment. And the result proved that convergence of Ant Colony Optimization is speeded up , and it has a strong global optimization capability.