Use this URL to cite or link to this record in EThOS:
Title: A Genetic Algorithm Approach to Analogue Integrated Circuit Placement
Author: Drakidis, Athanasios
ISNI:       0000 0001 3432 436X
Awarding Body: University of Essex
Current Institution: The University of Essex pre-October 2008
Date of Award: 2007
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
The explosive technological gro\V1h mainly driven by the telecommunications and infonnation technology nJarket necessitate the development of mixed-signal integrated circuits with their number of transistors constantly increasing. Such circuits typically contain analogue parts and large amounts of digital circuitry. The increasing complexity that encompasses every step in the design phase can only be overcome by the use of Electronic Design Automation (EDA). This requires mature methodologies that can in turn define clearly a suitable design flow via a complete set· of algorithms and procedures for solving th€./* complex problems arising, and thus minimising human intervention. It is generally agreed that these exist for the digital domain in which EDA allows the manufacture of circuits containing millions of transistors. In the analogue domain unfortunately a shortage of such tools exist. . This thesis focuses on the development of a Genetic Algorithm that targets the placement of circuit modules with an emphasis on analogue blocks and the constraints they impose. To tackle this problem considered, a step by step approach is taken that starts with a simple version of the placement problem (the digital placement problem) that is subsequently generalised and the algorithm is gradually extended. The result is an algorithm that is able to handle the placement of fixed and soft blocks for area/wirelength, and symmetry and proximity constraints often··arising in analogue circuits. The algorithm is equipped with a novel crossover operator that aiiows the use of other existing crossovers to be adapted and used for placement. In addition, to tackle the problem of defining the parameters of the algorithm, a selfadaptive methodology has been developed that allows some of these parameters to be included in the optimisation process along with the variables of the problem at hand. The algorithm is based purely on GA techniques and operators avoiding the use of hybrid methods such as local search operators that tend to be computationally expensive and slow down the optimisation process. The performance of the final tools is compared to other algorithms from the published literature and conclusions are drawn.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available