A Comprehensive Learning Cuckoo Search Algorithm For Solving Economic Dispatch Problems

Economic dispatch(ED)plays an important role in the economic operation of power systems.It aims to make a generation plan with minimum fuel cost for a given power demand while considering multiple practical constraints.Owing to the large-scale,non-convex benchmarks of the ED problem,traditional optimization methods are often trapped by local optimums and their solutions are not satisfying.Therefore,this paper introduces a swarm intelligence optimization algorithm,and manage to improve its performance as well as obtain a lower economic cost scheduling scheme by incorporating superior searching strategies and special constraint-handling mechanisms that based on the existing power dispatching knowledge.First,this paper designs an enhanced version of the cuckoo search algorithm,named comprehensive learning cuckoo search algorithm(CLCS),with two reinforcement strategies.A duplicate elimination strategy is designed to eliminate individuals of slow evolving efficiency using an elite searching strategy.It aims to prevent the issue of imbalanced evolving state and thus improve the convergence efficiency.Besides,a comprehensive-learning strategy is introduced to construct learning models using dimensional information.It enables every individual in the swarm of the algorithm learning high-quality experience from different individuals in each dimension,and thus improve the search ability of the algorithm.Then,CLCS is implemented to solve the ED problem.A chaos-Lambda method(CLM)method is proposed for generating initial swarm based on the equal incremental principle of classical power dispatching knowledge.It generates solutions in the neighbourhood of the ideal optimal solutions through the fitting formula of Lambda and power,which minimizes the search space and thus reduces the search complexity.Moreover,a slack approach is used to repair the solution that violates the constraint,and a CLM-based repair method is proposed to deal with unrepairable solutions.These methods can guarantee the feasibility of the solution under a large number of system constraints.Next,CLCS is introduce to solve the combined heat and power(CHP)economic dispatch(CHPED)problem.The CHP system is a hybrid energy system that coordinates the production and utilization of electric energy and thermal energy to improve energy efficiency.It is necessary to consider the supply and demand of both types of energies and satisfied their constraints.This paper introduces the proposed algorithm into the CHPED problem,and several effective mechanisms are designed for the bivariate constraints of electric and thermal energy.Finally,this paper selects several widely recognized models of power systems and CHP systems for numerical experiments,and the results are compared with state-of-the-art algorithms in terms of search ability,robustness and convergence speed.CLCS algorithm achieves the best performance in all tests.Mention that,in some large-scale cases,CLCS shows the potential to annually save millions of dollars.