Clonal Selection Algorithm Improvements And Its Application In3G Base Station Location

Artificial immune algorithm (AIA) is a kind of evolutionary computation methodbased on swarm intelligence, which is inspired by the human body immunology.Because the algorithm has advantages of simple model, fast convergence speed andeasy implementation, it has been widely used in the fields of scientific research andengineering applications. However, with the increasing complexity of the beingsolved optimization problem, and AIA itself also has some defects, such as lowconvergence accuracy, slow convergence speed in the later phase and prematureconvergence generated by easily falling into local optimal for solving the large-scalecomplex problems. Therefore, it has important theoretical significance and practicalvalue to do research on improving the algorithm and applying it to complexengineering optimization problems.The evaluation measure of the clonal selection in the basic Clonal SelectionAlgorithm (CSA) is the affinity of the antibody, it is easy to abandon potentialexcellent antibody, but these excellent antibodies can play a very important role in theglobal search. Based on this, a Backtracking Clonal Selection Algorithm formulti-modal function optimization (BCSA) is presented. The basic idea of this paperis as follow: using an improved backtracking mechanism and inhibit strategy ofmemory antibodies to maintain the diversity of antibodies and enhance the globalsearch ability. Improving the dynamic variation, selection and crossover operation tospeed up the algorithm convergence. The test results of typical multi-modal functionsshow that BCSA can effectively avoid falling into local optimum and accuratelysearch out the global extreme points of the multi-modal functions.Considering the poor performance of the basic CSA in optimizing complexfunction, we presented an improved-CSA on the basis of analyzing the principle andimprovement ideas, in order to overcome the defects such as premature convergenceof the algorithm and slow convergence speed in the later evolution phase in this paper.Firstly, the antibody initialization method is improved by the method of chaosopposition-based learning, with the result that the search space of the algorithm wasexpanded and the global convergence speed was improved. Then the method makesfull use of the strong exploration ability of Artificial Bee Colony (ABC) algorithmsearch operator to guide the population to jump out of local optima in the later stage of evolution to avoid premature convergence. The test results of12benchmarkfunctions and comparisons with other algorithms show that the algorithm has betterconvergence rate, higher precision and stronger global search ability.3G base station location is a key part of network construction. According to thedefects of the existing3G base station location optimization algorithm and thecharacteristics of the TD-SCDMA network, a optimization program of TD-SCDMAnetwork base station location based on immune algorithm was proposed in this paper.A problem description and a mathematical model of base station location wereestablished, a population initialization program based on opposition-based learningand a elite crossover strategy were designed, the immune optimization algorithmframework is presented. The experimental results show that the algorithm can notonly get higher network coverage with a relatively smaller consideration, namely thehigh cost-effective of the planning, but also has better convergence.