Paper
5 March 2018 Test scheduling optimization for 3D network-on-chip based on cloud evolutionary algorithm of Pareto multi-objective
Chuanpei Xu, Junhao Niu, Jing Ling, Suyan Wang
Author Affiliations +
Proceedings Volume 10710, Young Scientists Forum 2017; 1071013 (2018) https://doi.org/10.1117/12.2315901
Event: Young Scientists Forum 2017, 2017, Shanghai, China
Abstract
In this paper, we present a parallel test strategy for bandwidth division multiplexing under the test access mechanism bandwidth constraint. The Pareto solution set is combined with a cloud evolutionary algorithm to optimize the test time and power consumption of a three-dimensional network-on-chip (3D NoC). In the proposed method, all individuals in the population are sorted in non-dominated order and allocated to the corresponding level. Individuals with extreme and similar characteristics are then removed. To increase the diversity of the population and prevent the algorithm from becoming stuck around local optima, a competition strategy is designed for the individuals. Finally, we adopt an elite reservation strategy and update the individuals according to the cloud model. Experimental results show that the proposed algorithm converges to the optimal Pareto solution set rapidly and accurately. This not only obtains the shortest test time, but also optimizes the power consumption of the 3D NoC.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Chuanpei Xu, Junhao Niu, Jing Ling, and Suyan Wang "Test scheduling optimization for 3D network-on-chip based on cloud evolutionary algorithm of Pareto multi-objective", Proc. SPIE 10710, Young Scientists Forum 2017, 1071013 (5 March 2018); https://doi.org/10.1117/12.2315901
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Network on a chip

Clouds

Evolutionary algorithms

Multiplexing

Optimization (mathematics)

3D modeling

Genetic algorithms

Back to Top