An Improved Ant Colony Optimization And Its Application In The Power System

With the continuous development of the electric power industry，the scheduling problemof the power system is more and more important. In which, the main goal of the power systemscheduling is to control the entire operation of the power system, and ensure the responsibilityof accurate and high-quality power supply, no matter in normal situation or in the accidents.This goal means that the system should guarantee the quality, the economy and the safety ofwhen supplying power, and can manage the accidents reasonably when facing problems.The power system is a very complex system, and there are many nonlinear combinatorialoptimization problems in it. In recent years, many algorithms have been used in the powersystem, including the genetic algorithm, the colony algorithm and many other algorithms, inorder to improve the quality and efficiency of power generation and ensure the economicbenefits.Ant colony optimization (ACO), is a bionic method of evolution, it was proposed byItalian scholar Marco Dorigo in the European Artificial Life Conference in1991,heintroduced the inspired origin and the basic model of the ant colony algorithm in”Distributedoptimization by ant colonies”, then he explained the idea of the algorithm in his paper in1992.This algorithm has many advantages, including robustness, easy to combine with otheralgorithms, the distributed computing and so on. So far, this algorithm has been successfullyused in many fields, including the traveling salesman problem(TSP), resource quadraticassignment problem (QAP),the integrated circuit network layout, communication networkrouting, the power system and so on.Although this system has many advantages and has been used in many problems, due tothe time that this algorithm raised is relative late, there are some shortages in the algorithm,such as premature, the long time for search, and easy to stagnation and so on. For the defects, based on basic algorithm, we improve it. We improve the initialization,the selection probability and the update strategy based on MMAS, then we apply thealgorithm into the power system and analyze the results, finally we prove its advantage.