Circuit Testing Based On Ant Colony Algorithm Research

With the rapid development of IC industry, the requirement of circuit testing iskeeping increasing. Consequently, the research of digital circuit testing is full oftheoretical and practical significance.First of all, the principle of random testing method is introduced. The inner circuitmodel is constructed using C++, then the single stuck-at fault is simulated. Therandom testing method is realized in computer program.Secondly, the ant system is used to study the best collocation of 2-dimensionalhybrid cellular automata that can generate test vectors of good random. The topologyof cellular automata and the structure rule of each cell are designed to generaterandom vectors with better quality. Aiming at using cellular automata to producetesting vectors of better quality, the regulation of max-min ant system is fine-tuned tooptimize the collocation of cellular automata. The circuit testing experiments showthat the ant system can find the best collocation of cellular automata, and the cellularautomata optimized by max-min ant system can generate pseudorandom numberarrays of good random, also the cellular automata can be applied to digital circuittesting to improve the fault coverage and the test length.Lastly, the multilevel thresholding method of graphics which is based on 2-dimensional histogram is presented. The multiple regionalization of 2-dimensionalhistogram and related 2-dimensional entropy are studied. The fast iterative algorithmfor calculating 2-dimensional entropy is induced. The regulation of ant system isadjusted to optimize the multiple regionalization of 2-dimensional histogram, in orderto realize the multithresholding of graphics. One the whole, the graphics segmentationexperiments achieved high performance.