In this paper an application of Cellular Learning Automata (CLA) to VLSI placement is presented. The CLA, which is introduced for the first time in this paper, is different from standard Cellular Learning Automata in two respects. It has input and the cell neighborhood varies during the operation of CLA. The proposed CLA based algorithm for VLSI placement is tested on number of placement problems and has been compared with several reported algorithms such as: simulated annealing, genetic algorithms, the algorithm proposed by Saheb Zamani and Hellestrand, and the algorithm based on Kohenen neural network. The results obtained show that the proposed algorithm produces results, which are comparable to the other algorithms reported in the literatures. The parallel nature of CLA makes it appropriate for hardware implementation.

Keywords: Learning Automata, Cellular Learning Automata, Placement, Very Large-Scale Integrated circuits, Gate Array

Full-Text [PDF 2048 kb]   (2516 Downloads)