Improvement Of Wolf Colony Algorithm And Its Application In Complex Function Optimization Problems

According to some evolutionary phenomena in the natural world and the group characteristics of the organisms,some intelligent optimization algorithms are constructed and used to solve complex function optimization problems.Wolf colony algorithm was first proposed in 2011 as a new swarm intelligence optimization algorithm,which was designed based on the predation behavior of the wolves.Now,it has been applied in many fields such as optimization of three-dimensional sensors,backpack problems,drone planning and optimal operation of hydropower station reservoirs,and good results have been achieved,which makes it becomes one of the group intelligent optimization algorithms with broad application prospects.However,there are some shortcomings when using this algorithm to solve some practical problems,such as the accuracy of solution,the speed of convergence,the application area of the extended algorithm,etc.Based on the basic wolf colony algorithm,this thesis proposes two improved algorithms,and applies it to specific examples.The main research content is as follows:1.Based on the basic wolf colony algorithm,an improved wolf colony algorithm based on adaptive step size is presented(Adaptive Step Wolf Colony Algorithm,ASWCA),and the location problem of logistics distribution center is solved by using this algorithm.In the basic wolf colony algorithm,the attack step in the summon behavior and the siege step in the siege are fixed,which affects the optimization performance of the algorithm.Therefore,the nonlinear dynamic inertia weight coefficient formula is adopted in the attack step,which makes the value of the attack step automatically adjusted by the change of the fitness value,thus the intelligence is increased in the search process.The siege step further adopts the adaptive updating formula,which reduces the siege step with the increasing number of iterations,thus the probability of finding the better value is improved.Then,through the simulation of the classical test function,the results show that the proposed algorithm has a good global search ability.That is,the optimization performance of the algorithm is improved.Finally,the proposed algorithm is applied to solve the location problem of logistics distribution center,and good results are achieved.2.A new wolf colony algorithm(New Wolf Colony Algorithm,NWCA)is introduced and applied to solve traveling salesman problem by introducing disturbance operation and Sigmoid function into basic wolf colony algorithm.Because the basic wolf colony algorithm has great blindness and randomness in the search process,As a result,the algorithm is difficult to find the global optimal solution in a short time.So,first of all,in the walk behavior of the algorithm,the perturbation operation is integrated,and the individual components on the position of the wolves are minutely adjusted to make the wolves fine search within a very small range.Then,in the calling behavior,we use the Sigmoid function to construct the attack step length to make the attack step size decrease gradually in the prescribed range.Next,in the algorithm's siege behavior,we add regulatory mechanism to enable wolves to have the ability to regulate food when they are besieged.Finally,the algorithm is verified to be effective by simulating the typical test functions and solving the traveling salesman problem.