GMMSO: game model-combined improved moth search optimization approach for reconfigurable asymmetric multi-processor system-on-chip architecture. Issue 4 (3rd April 2023)

• Record Type: Journal Article
• Authors: Mariyappan, Isaivani; Veluchamy, Malathi
• Is Part Of: Engineering optimization. Volume 55:Issue 4(2023)
• Page Start: 685
• Page End: 702
• Publication Date: 2023-04-03
• Subjects: Very large-scale integration -- multi-processor system-on-chip -- optimization -- game model -- moth search algorithm -- field programmable gate array; Engineering design -- Periodicals; Mathematical optimization -- Periodicals; 620.0042
• DOI: 10.1080/0305215X.2021.2024815
• Languages: English
• ISSNs: 0305-215X

Abstract

Abstract : Heterogeneous Multi-Processor System-on-Chip architectures are prevalent in modern embedded system applications that target high-performance needs. This work comprises two parts to optimize the multiprocessor system structure concerning system speed, power, and area requirements. The first part includes the game model concept, which is remarkably advantageous in generating strategic decisions on incoming signals and minimizes traffic delays coming from the input portion of the system. Secondly, an improved moth search optimization is proposed to explore the design space and optimize the system space and power consumption. This research is implemented in the Field Programmable Gate Array (FPGA) Xilinx Zynq and Virtex platforms and evaluated by ExPRESS benchmarks. The Game Model-combined improved Moth Search Optimization (GMMSO) technique showed performance improvements of 69.34% in delay measurement, 46% in data rate and 36.34% in static power consumption compared with the traditional HMILP (Heuristic Mixed Integer Linear Programming) methods.