Research On Improved Artificial Bee Colony And Its Application In Multiuser Detection

Artificial bee colony algorithm is a new swarm intelligence algorithm inspiredby the intelligent foraging behavior of honeybee swarms. It realizes swarmintelligence through the exchange, transformation and cooperation between thedifferent roles of bees. Because of its less control parameters, ease of implementation,simplicity and robustness, it has been concerned by more and more scholars andsuccessfully applied to various fields such as function optimization, filter design,image processing and wireless communications. Optimum multiuser detection canprovides the minimum bit error rate, the highest asymptotic efficiency and the bestanti-near-far effect capacity. However, it has been proven that the optimum multiuserdetection is a non-deterministic polynomial (NP) problem. As a simple, effective andnew heuristic algorithm, artificial bee colony algorithm can effectively solve thisproblem. Therefore, the combination of artificial bee colony algorithm and optimummultiuser detection has important theoretical significance and practical value toimprove the system capacity, and enhance system performances.The main work is summarized as follows:(1) A differential evolution binary artificial bee colony algorithm was proposed.A neighborhood search formula directly in discrete domain was designed.Multidimensional neighborhood search strategy was used and algebraic operationswere replaced by logical operations. The improved algorithm’s convergenceperformance was analyzed and the effectiveness was verified through the0-1knapsack problem’ simulation. And a multiuser detection algorithm based ondifferential evolution binary artificial bee colony was proposed. The simulation resultsshow that the performances of the improved algorithm are better than that of thetraditional algorithm in convergence property, anti-multiple access interference andanti-near-far effect.(2) A particle swarm binary artificial bee colony algorithm was proposed. Globaloptimal solution information was used to guide the generation of candidate solutionsand the improved neighborhood search formula was proposed. The convergenceperformance of the improved algorithm was analyzed. The test function’s simulationdemonstrates the superiority of the algorithm. The improved algorithm was applied tooptimize the objective function of multiuser detection technology, and a multiuser detection technology based on particle swarm binary artificial bee colony algorithmwas proposed. The objective function was designed and the algorithm was simulated.(3) A distribution estimated binary artificial bee colony algorithm was proposed.Food source update formula combined global group information and local individualinformation was designed. It utilized the statistics information of global high-qualitysolution obtained from estimation of distribution algorithm to generate candidatesolution. The analysis of convergence performance and the simulation experimentwere carried out. Meanwhile, a multiuser detection algorithm based on distributionestimation binary artificial bee colony was proposed. The convergence properties,anti-multiple access interference and anti-near-far effect performance were verifiedthrough computer simulations.