Study On Optimization Methods Based On Biological Swarm Intelligence And Applications

The ecosystem is a source of inspiration for human science and innovation. The nature of biological intelligent behavior is the origin, volutionary history and ultimate destination of the artificial intelligence research. The natural life is subtle and perfect, while the group that made up of individuals demonstrate more splendid scene. Birds are able to synchronize flight without centralized control; a group of bees can create beautiful honeycomb; minimum single.cell biological bacteria, can find the shortest path in a maze, or connect different arrays of food in an efficient manner, but no fault tolerance. That is, in long course evolution of the natural world, all life forms, after a brutal survival of the fittest, finally are able to survive, and have the extraordinary ability to survive and wisdom. Therefore, inspired by the robustness and adaptability of natural system in response to the complex environment, the researchers of biological systems raised many computational models and algorithms that mimic the biological behavior to solve complex engineering optimization problems. These intelligent optimization methods based on the biological behavior illustrated many advantages of a wide range of applications, high optimize performance and efficient, no special information, and robustness. Nowdays, the biological inspired methods have been applied in many practical applications and show great potential for development.This work performs depth study of several biological heuristic computation techniques from the concept, nature, models, and methods aspects. Combined with hot, difficult, and key issues in the current biological heuristic computation, this work carried out profound research from two aspects of the theory and engineering applications, and has made a lot of innovative and valuable research results as follows:(1) Improved bee colony algorithm studies and its applications On the basis of analysis of the basic ABC algorithm, the discrete version of the ABC algorithm (BABC) and the improved algorithm based on the structure of information exchange (TABC.V). The BABC algorithm to fill the gaps in the research of classical ABC algorithm for solving discrete optimization problems; by embedding the information exchange topology into classical ABC, the TABC.V algorithm significantly improves the optimize performance of ABC algorithm. Through the simulation tests on Discrete and continuous functions, it is verified that the two improved algorithms can efficiently solve these optimization problem, by overcoming the premature convergence problem of the basic algorithm. Then the BABC and TABC.V algorithms were applied to solve the knapsack problem and neural network training. The simulation results show that the two improved algorithms have excellent engineering optimization problem solving ability.(2) Symbiotic coevolutionary optimization model and its application to RFID networks schedulingThis work proposed a symbiotic multi.colony coevolution optimization model by combining co.evolution theory and nature of biological symbiosis. By embedding the standard PSO algorithm to the multi.colony coevolution model, a multi.colony symbiotic co.evolution of PSO algorithm (MSPSO) is proposed. According to the three typical symbiotic modes, three versions MSPSO algorithm, namely MSPSO.C, MSPSO.P, and the MSPSO.M is developed. The simulation results show that the MSPSO algorithm can rapidly convergence to the global optimal solution of the problem, while maintaining the diversity of the population. That is, MSPSO gains the potential for solving complex engineering optimization problems. Then the MSPSO algorithm was applied to solve large-scale RFID network reader scheduling problem. Compared to PSO and GA, simulation study on four different scale RFID networks showed that the MSPSO algorithm has a clear advantage to efficiently solve the scheduling problem of large.scale RFID networks.(3) Research and application of novel bacterial colony foraging algorithm based on optimal foraging and social learning theoryOn the basis of existing bacterial optimization algorithm models, the introduction of bacterial adaptive feeding mechanism and quorum sensing mechanism, flora foraging optimization algorithm (BCF). BCF algorithm will be applied to solve new synthetic function optimization problems, and other traditional swarm intelligence optimization algorithm performance, the results show that the BCF has a complex function optimization problem solving ability. BCF algorithm used in data mining, clustering analysis of the problem, BCF.based clustering analysis algorithm, the performance of typical data clustering results showed that the biological heuristic algorithm based on clustering analysis algorithm with the existing success of BCF compared to faster convergence, and the clustering of higher quality.