High performance genetic algorithm for VLSI circuit partitioning

Simona Dinu

doi:10.1117/12.2243196

14 December 2016 High performance genetic algorithm for VLSI circuit partitioning

Simona Dinu

Proceedings Volume 10010, Advanced Topics in Optoelectronics, Microelectronics, and Nanotechnologies VIII; 1001029 (2016) https://doi.org/10.1117/12.2243196
Event: Advanced Topics in Optoelectronics, Microelectronics, and Nanotechnologies 2016, 2016, Constanta, Romania

Abstract

Partitioning is one of the biggest challenges in computer-aided design for VLSI circuits (very large-scale integrated circuits). This work address the min-cut balanced circuit partitioning problem– dividing the graph that models the circuit into almost equal sized k sub-graphs while minimizing the number of edges cut i.e. minimizing the number of edges connecting the sub-graphs. The problem may be formulated as a combinatorial optimization problem. Experimental studies in the literature have shown the problem to be NP-hard and thus it is important to design an efficient heuristic algorithm to solve it. The approach proposed in this study is a parallel implementation of a genetic algorithm, namely an island model. The information exchange between the evolving subpopulations is modeled using a fuzzy controller, which determines an optimal balance between exploration and exploitation of the solution space. The results of simulations show that the proposed algorithm outperforms the standard sequential genetic algorithm both in terms of solution quality and convergence speed. As a direction for future study, this research can be further extended to incorporate local search operators which should include problem-specific knowledge. In addition, the adaptive configuration of mutation and crossover rates is another guidance for future research.

Citation Download Citation

Simona Dinu "High performance genetic algorithm for VLSI circuit partitioning", Proc. SPIE 10010, Advanced Topics in Optoelectronics, Microelectronics, and Nanotechnologies VIII, 1001029 (14 December 2016); https://doi.org/10.1117/12.2243196

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available