Home         Authors   Papers   Year of conference   Themes   Organizations        To MES conference

Problem of a transistor length variation in a standard cell at multi-objective optimization of nano-size VLSI

Authors
 MelikAdamyan A.F.
 Ryzhov A.P.
Date of publication
 2010

Abstract
 With shrinking CMOS technology, the accurate trade-off between delay, static power consumption and yield of a digital circuit is becoming the most important factor while designing a functionally reliable and low power circuit. Gate sizing has emerged as one of the efficient way to achieve the goal in post-layout step of design flow. In the past single-objective optimization techniques have been used to optimize the timing variation, power or yield whereas on the other hand multi-objective optimization techniques can provide a more promising approach to design the circuit. We propose an algorithm based on multi-objective optimization technique called Non-dominated Sorting Genetic Algorithm. It optimizes a circuit in multi domains and provides the user with Pareto-optimal set of solutions which are distributed all over the optimal design spectrum, giving users the flexibility to choose the best fitting solution for their requirements. Algorithm overcomes the disadvantages of traditional optimization techniques.
Keywords
 CAD of microelectronicx, genetic algoriths, multi-objective optimization
Library reference
 MelikAdamyan A.F., Ryzhov A.P. Problem of a transistor length variation in a standard cell at multi-objective optimization of nano-size VLSI // Problems of Perspective Micro- and Nanoelectronic Systems Development - 2010. Proceedings / edited by A. Stempkovsky, Moscow, IPPM RAS, 2010. P. 204-207.
URL of paper
 http://www.mes-conference.ru/data/year2010/papers/m10-290-61011.pdf

Copyright © 2009-2019 IPPM RAS. All Rights Reserved.

Design of site: IPPM RAS