Bionic Intelligent Computing

In this thesis, we mainly investigate the Biologicallyinspired Illtelligent Com-puting Methods. After eons of evolution, nature fascinates us with its extraordinaryefficiency of performing various tasks. Probing and simulation its mechanism not onlybenefit us for numerous kinds of application, but most probably, bring us more insightinto the basic questionwhat is intelligence.Here, we focus on four topics f ArtificiaJ Neural 3:etwork(ANN), Swarm Intelli-gence(SI), EvolutiOnary AlgorithJn(EA) and DNA Computing. Rom microscopic DNAto macroscopic insect swarm, the foux explore intelligence from differeot perspectivesat different scales.. \ANN simulates the structure of our brain and expect intellig6nce from the networkof neuxons. First we introduce its basic idea, tyPical activating functions, learning rulesand its main aPplication: Classification &C1ustering, associated memory and optimization. We then focus on our proposed Alternate Covering Neural Network(ACNN) andits application in "DNA Sequence Classification" and "UCI WaVeform Data Classi-fication". The first application (DNA Sequence Classification) is presented in detailwith the process of feature eXtracting. Furthermore, we discuss the statistical learningtheory and the equivaence of Support Vector Machine with threelayered feedfOrwardneuraI network.SI aims at understanding the collective behavior of swarm insects. We browsebriefly its concepts, characteristics and major research areas:Sorting behavior, AntColony Optimzation(ACO) and Particle Swarm Optimization(PSO). We then presentour mathematicaJ model of nest building using randomlyconnected neural network.Proof is made with theory of Iterated Function System(IFS)* ExPerimental resuJts ofsforlation are also presented.EA simulates the "Nature Selection" and genetic phenomenon and applies to opti-mization application successfully. We here mainly investigates the genatic aJgorithm,especially our good-point based g6netic algorithm(GGA). Applying GCA to Job-shopTTScheduling problem is presented.DNA Computing is a new-born area with great potential for new-tyPe computingother than electronic and quatum computing. We here introduce its origin, basic ideaand application in Haxnilton Route problem, and make some comment.