Title :
Application of evolutionary algorithm to three key problems in VLSI layout
Author :
Nan, Guo-fang ; Li, Min-qiang ; Lin, Dan ; Kou, Ji-Song
Author_Institution :
Inst. of Syst. Eng., Tianjin Univ., China
Abstract :
Evolutionary algorithm is a class of stochastic search algorithm, which can be applied to both combinatorial and numerical optimization problems, especially NP hard problems. Circuit partitioning, placement and clock routing are three key phases in VLSI physical design and they are proved to be NP hard. So a genetic algorithm frame and its realization process are presented in this paper in order to solve these problems. Meanwhile, this algorithm is used to test different benchmarks for three different problems, experimental results show that it is a feasible and valid algorithm. This genetic algorithm can also improve solutions when compared with traditional heuristic methods.
Keywords :
VLSI; computational complexity; electronic engineering computing; evolutionary computation; integrated circuit layout; stochastic processes; NP hard; VLSI layout; VLSI physical design; circuit partitioning; circuit placement; clock routing; combinatorial optimization; evolutionary algorithm; genetic algorithm; numerical optimization; stochastic search algorithm; Benchmark testing; Circuit testing; Clocks; Evolutionary computation; Genetic algorithms; NP-hard problem; Partitioning algorithms; Routing; Stochastic processes; Very large scale integration; Evolutionary algorithm; NP hard; genetic algorithm; physical design;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527443