Research On Co-evolution Algorithm Based On Quantum Particle Swarm

Quantum-behaved particle swarm optimization(QPSO)is a novel swarm intelligence optimization algorithm based on quantum mechanics,which is an improved version of particle swarm optimization(PSO).Due to the randomness of quantum,the algorithm has the advantages of global search so that it has excellent performance in the optimization problem.At present,quantum-behaved particle swarm optimization has attracted extensive attention from scholars both at home and abroad.It has been widely applied in many applications,such as finance,clustering and classification,neural network,signal processing and so on.However,the diversity of particle swarm is inversely proportional to the number of iterations of the algorithm in the QPSO algorithm.With the increase of the number of iterations,the diversity of particle swarm is reducing,the global search ability of the algorithm is weakening,the algorithm can get into local optimal in the later stage of evolution in the end.In order to improve the defects of the QPSO algorithm which is easy to fall into local optimum,this paper proposes a multiple groups mechanism,a group scoring function mechanism and a cooperative mechanism under the inspiration of co-evolution algorithm to improve the performance of the algorithm.Finally,this paper propose a cooperative quantum particle swarm optimization based on multiple groups(CGQPSO).In the CGQPSO algorithm,a multiple group mechanism which a group is composed of several particles as a basic unit of evolution is introduced to give full play to the swarm intelligence.The scoring function mechanism is introduced to evaluate the searching ability of the group which includes the global search capability and the local search ability.The collaborative mechanism based on inter species competition relationship is introduced to adjust search strategy dynamically and adaptively in order to reduce the dependence of algorithm on optimization problem and balance global search ability and local search ability of algorithm.Finally,experimental analysis is carried out to evaluate the performance of algorithm,convergence speed of algorithm,stability of algorithm and time complexity of algorithm.It shows that the algorithm proposed in this paper has the characteristics of high performance,high stability and no easy to fall into local optimum.In order to further apply the multiple group mechanism,the group scoring function mechanism and the cooperative mechanism into 0-1 knapsack problems and multidimensional knapsack which are combinatorial optimization problem.Firstly,the discrete binary quantum particle swarm optimization algorithm is proposed by discrete processing of the QPSO algorithm.Then,because of the solution space of knapsack problem is discrete,this paper constructs the group scoring function based on Hamming distance.Finally,This paper proposes a discrete quantum particle swarm optimization algorithm based on cooperation for knapsack problem.Experiments which are tested on 0-1 knapsack problems and multidimensional knapsack problems show that the proposed algorithm has high accuracy in solving discrete problems such as knapsack problem.