Improved Bat Algorithm And Its Application In Cognitive Radio Spectrum Allocation

The best solution has always been the goal that people are pursing in the project.However,with the problems that people studied become more and more complex,constraints and large scale,finding the optimal solution becomes more difficult.Considering dynamic programming and combinational optimization problem,the traditional methods can not get the optimal solution in effective time because of the increasing solution space scale.However,meta heuristic algorithm has a unique advantage in solving these problems.Bat algorithm is a kind of meta heuristic algorithm,which has less parameters,simple structure,fast convergence speed,good balance of local search and global search and other advantages,and has been widely used in many fields.In this paper,in order to improve the optimization performance of original bat algorithm for its shortcomings,the algorithm is analyzed and researched deeply.Mainly from the following aspects to improve:(1)The original bat algorithm changes the step size of the bat to the current optimal solution by frequency.However,in the improved bat algorithm,different strategies are chosen according to the distance between the bat and the current optimal position in the population.If the bat is far from the current optimal solution,the bat flies to the optimal position by random step size.If the bat is closed to the current optimal solution,it suggests that the bat is nearby the current optimal position,hence,the bat flies randomly.(2)Through the analysis of the trajectory of bats,it is found that many bats are trapped into local optimal point because of decreasing diversity of bats.Other bats are directed only by current optimal solution,and there are no mechanisms to jump out of local optima.Therefore,the behavior of quantum of bats is introduced into the algorithm to increase the diversity of population and it also contributes to avoiding prematuration.(3)During the search time,the position of each bat is not only updated by the current optimal solution,but also depends on the mean best position.The mean best position is the average value of the best positions that bats experienced in the population,which makes the bat far away from the best position to fly to its vicinity.That helps to speed up the convergence speed of the algorithm.In order to verify the effectiveness of the improved bat algorithm,the optimization ability of the algorithm is tested by the benchmark functions.The simulation results show that the improved bat algorithm can effectively improve the search precision and speed up the convergence speed of the algorithm.Since the problem of spectrum allocation is a discrete domain optimization problem,it is necessary to discretize the bat algorithm so that it can be applied in spectrum allocation of cognitive radio.According to the spectrum allocation based on graph theory,the optimization results of discrete bat algorithm(binary bat algorithm)are not good.Therefore,the binary bat algorithm is changed to improve the benefit value of system.Mainly from the following aspects to improve:(1)The selection strategy of the bat algorithm is changed into greedy selection,which enhances the exploit ability of binary bat algorithm in the current position.(2)By analysizing the distribution of best position that each bat experienced,the statistic characteristic of bat's position was used to guide the bats finding the optimal solution.It can enhance the speed of convergence of bat algorithm.During the local search,the solutions are't obtained by mapping the real number to binary set but updated directly in the discrete domain,which shorten the searching time.The improved algorithm is applied to spectrum allocation,and the simulation results are compared with other algorithms.The simulation results not only show that the proposed algorithm is more efficient,but also show the speed of convergence is faster than the other algorithm.