Premium
Parallel computing modes of multiobjective game
Author(s) -
Bao Wei,
Pan Chuang Ye,
Wang Lu,
Zhang Qing,
Xie Neng Gang
Publication year - 2020
Publication title -
concurrency and computation: practice and experience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.309
H-Index - 67
eISSN - 1532-0634
pISSN - 1532-0626
DOI - 10.1002/cpe.5687
Subject(s) - computer science , speedup , mode (computer interface) , block (permutation group theory) , disadvantage , mathematical optimization , process (computing) , parallel computing , distributed computing , mathematics , artificial intelligence , geometry , operating system
Summary For the design of multiobjective optimum decision, three methods based on parallel computing and game theory were raised to transfer the multiobjective design into the decision. Then, the multiobjective optimum design problems of numerical calculation examples, compensative sheave block, and antiseismic design of arch structures can be solved. Three parallel computing modes (ie, I, II, and III) were designed correspondingly. First, parallel mode I can make the game process visually intuitive with the disadvantage that the computing load cannot be balanced. Second, parallel mode II can better solve the imbalance of computing load between different objectives, obtain good parallel speedup and host efficiency. Furthermore, parallel mode III has the advantages of the above two parallel modes. However, its disadvantage is that data exchange between computing units is frequent, and this results in an increase in communication time. The simulation results suggest that all the parallel game algorithms can give considerations to multiple objectives, improve the seeking efficiency and make the solutions more robust.