Research On Artificial Bee Colony Algorithm And Its Application In The Spectrum Sensing Of Cognitive Radio

Artificial bee colony (ABC) algorithm is a swarm intelligence optimization algorithmwhich imitates the behavior of honey bee. Through mutual communication, role reversaland cooperation between different roles bee achieve the intelligence. The ABC algorithmhas the characteristics of less control parameters, simple calculation, easy realization, andstrong robustness and so on. It is successfully applied in function optimization, imageprocessing, filter’s design, blind signal separation, wireless communications and artificialneural network. ABC algorithm research is in its infancy, and there are still many issues tobe resolved. For example, for some complex objective functions, convergence accuracy isnot high; in later iterations the reducing of population diversity may lead to slowconvergence.On the basis of overview and analysis of the original ABC algorithm, this papersummarizes the main features and problems of the algorithm. Then the improvement ofABC algorithm is proposed. For the problem of population diversity reducing, the conceptof the average distance amongst points and population distribution of entropy isintroduced as the adaptive disturbance condition, to improve the population diversity; forcomplex optimization problems, convergence accuracy of ABC algorithm is not high,adaptive correction factor is introduced to adjust the worst particles in the population, andimprove the convergence precision; for low convergence speed of the ABC algorithm, thescouts taking full advantage of the historical information and searching in the vicinity ofthe optimal solution can increase convergence speed. Combining the above three aspects,by introducing adaptive mutation mechanism, guaranteed convergence mechanism andoptimal tracking mechanism, an effective adaptive artificial bee colony (EA-ABC)algorithm is proposed. Simulation results illustrate the effectiveness of the EA-ABCalgorithm.Finally, the EA-ABC algorithm is applied to cooperative spectrum sensing for acognitive radios field. The cooperative spectrum sensing problem is reformulated into aconstraint optimization problem, and the EA-ABC algorithm is used to solve theoptimization problem. Then simulations are conducted to compare the performance of the proposed cooperative spectrum sensing method based on the EA-ABC algorithm with thatof the cooperative spectrum sensing method using the ABC, particle swarm optimization(PSO), and modified PSO algorithms. The simulation results validate the effectiveness andreliability of the proposed method and demonstrate that EA-ABC-based can achievehigher detection probability than other methods under the same false alarm probability.