An Evolutionary Approach for Pareto-optimal Configurations in SOC Platforms

: One of the most important problems in SOC platforms design is that of defining strategies for tuning the parameters of a parameterized system so as to obtain the Pareto-optimal set of configurations that provide multi-criteria optimisation. The paper proposes a methodology based on evolutionary techniques for exploration of the range of possible configurations of a parameterized system [1]. A highly parametric system-on-a-chip for digital camera applications will be taken as a case study and a multi-objective genetic algorithm will be used to search for the power-performance trade-off surface. The methodology proposed will be compared with that implemented in Platune [2] in terms of both accuracy and efficiency in relation to the number of simulations performed. Key words: Parameterized systems, system-on-a-chip architectures, design space exploration, genetic algorithms, multi-objective optimization, Pareto-optimal configurations, power/performance-tradeoffs.

• Type: Conference
• Year: 2001
• Where: IFIP
• Authors: Giuseppe Ascia, Vincenzo Catania, Maurizio Palesi
• Keywords: Genetic Algorithms, IFIP 2001, Multi-objective Genetic Algorithm, Programming Languages, SOC Platforms Design