Research On Deployment Of RFID Network Based On Improved Particle Swarm Optimization

Radio Frequency Identification technology(RFID) provides a fast general purpose dataacquisition and transmission for many industries. RFID system optimization models has thecharacteristics of non-linear and multi-objective, it is difficult to achieve satisfactory resultsusing conventional optimization algorithms for processing. Complex optimization can besolved by intelligent optimization methods, among which particle swarm optimization(PSO)attracts wide attention and application. PSO algorithm is simple, easy to implement and hasstrong robustness, very suitable for solving complex optimization problems, but it is easy tofall into local optimum and its convergence accuracy is low. Therefore, research andimprovement are carried out to optimize its optimization performance, then the improvedalgorithm is applied to the deployment of RFID network optimization problems.First the RFID system components and working principle are introduced, based on thedescription of the problem RFID network deployment, propagation models and readerantenna performance, consider the network deployment tag coverage, signal strength, networkanti-collision, intensity, load balancing, and crosstalk level of six factors, and to establish thecorresponding objective function, then classfy the objective functions into maximized andminimized optimization problems.Second, the label coverage and the level of cross-interference of two factors areconsidered, optimization objective function is established to meet maximized optimizationproblem. An objective function optimization with Tent mapping is introduced to improve thestandard particle swarm optimization. The algorithm parameters are installed by chaos, andchaotic variability is proposed to improved the global search ability of the algorithm when itfalls into local optimum. In order to illustrate the effectiveness of the Tent map and identifyalgorithm parameters for the case study, two simulation experiments are designed in this paper.(1)Performance comparison between Tent mapping and the most widely used Logisticmapping,(2)Uniform method of design is used to determine the optimal combination ofparameters. Tent_PSO with the optimal combination of parameters is used for RFID networkdeployment, and better effect is achivedFinally, four factors: the network signal strength, the intensity of the conflict constraints,intensity level and load balancing is considered, then an objective function is established inline with the minimization optimizing problem, meanwile elite opposition-based learningparticle swarm optimization is proposed when the quantum particle swarm optimizationfalls into local optimum (EOBL_QPSO). Four classic test function test is used to test theperformance of EOBL_QPSO, and comparative analysis proved the effectiveness of theimproved algorithm. Comparative simulation experiments show that EOBL_QPSO performwell in RFID solving multi-objective optimization problem of RFID network deployment, theconvergence of the algorithm is rapid, and the optimal deployment solution can be solvedeffectively.