作者: Taj-ul Islam
DOI:
关键词: Stochastic neural network 、 Benchmark (computing) 、 Routing (electronic design automation) 、 Rate of convergence 、 Interval graph 、 Artificial neural network 、 Computer science 、 Algorithm 、 Cellular neural network 、 Computer network 、 Time delay neural network
摘要: Neural network architectures are effectively applied to solve the channel routing prob lem. Algorithms for both two-layer and multilayer channel-width minimization, constrained via minimization proposed implemented. Experimental results show that algorithms much superior in all respects compared existing algorithms. The optimal solutions most of benchmark problems, not previously obtained, obtained first time, including an solution famous Deutch's difficult problem. four-layers one be lchmark is time. Both convergence rate speed with which simulations executed outstanding. A neural problem also presented. In addition, a fast simple linear-time algorithm presented, possibly coloring vertices interval graph, provided line intervals given.