Method Of Unit Commitment Based On An Improved Artificial Bees Colony Algorithm

Unit commitment problem is a typical optimization problem in power system, which determines the running status of units and the output power for every period in a scheduling cycle of the future in the condition of meeting the load of power system, reserving requirements and unit operating constraints of the technical conditions in order to minimize the total running costs of power system. Unit commitment problem is a typical large-scale nonlinear mixed integer programming problem, with the characteristics of high-dimensional, discrete and non-linear.This paper firstly describes the basic principles of artificial bee colony algorithm, including biological models, algorithm theory, algorithmic process and the strengths and weaknesses of the algorithm. Then the artificial bee colony algorithm is applied to the study of unit commitment problem. The paper separates the unit commitment problem into two sub-problems. They are determing the unit status and economic dispatch. The following improvements are proposed according to the characteristics of unit commitment problem:For the traditional unit commitment problem, the memory mechanism of immune algorithm is introduced into artificial bee colony algorithm which can avoid premature and falling into local optimum in order to improve solution efficiency through retaining the excellent unit pieces by memory bank; In the exploitation process, rank fitness selection strategy is used which shows better robustness on the basis of maintaining the diversity of individuals; In the exploration process, role play neighborhood searching method is used which maintains the high-quality characteristics of the original solutions and improves the searching speed.Subsequently, the improved artifical bee colony algorithm is applied in the problem of security constraints unit commitment with wind farm and I make further improvements for the model:I propose apiecewise linear method for objective function in order to accelerate the speed of optimization; By the results of solving the linear model, heuristic classification is proposed and reduces the optimization space of unit commitment and the size of the calculation. For the fluctuations and randomness of the wind output power, interval stochastic simulation method is proposed, which combines the randomness and the regularity of the forecast, and can minimize the cost of conventional generators and blindness can be avoided.In this paper, several examples are tested to validate and summarize the feasibility and effectiveness of improved artificial bee colony algorithm, and finally, the lack of this paper and the orientation of endeavor in the future are proposed in (?)nd.