CAJ | 학술논문

随着片上网络集成度的提高，低功耗映射逐渐成为研究的热点。文中利用调整序思想，将混合蛙跳算法应用到片上网络低功耗映射问题中；同时对混合蛙跳算法进行改进，引入自适应混沌方法优化较差个体，采用多邻域退火策略优化较优个体。实验结果表明，相比于传统混合蛙跳算法和模拟退火混合蛙跳算法，改进混合蛙跳算法具有更优的搜索能力，从而能进一步降低通信功耗。
수착편상망락집성도적제고，저공모영사축점성위연구적열점。문중이용조정서사상，장혼합와도산법응용도편상망락저공모영사문제중；동시대혼합와도산법진행개진，인입자괄응혼돈방법우화교차개체，채용다린역퇴화책략우화교우개체。실험결과표명，상비우전통혼합와도산법화모의퇴화혼합와도산법，개진혼합와도산법구유경우적수색능력，종이능진일보강저통신공모。
With the density of network-on-chip(NoC) integration getting higher,the low-power mapping has become a hotspot research.A novel shuffled frog-leaping algorithm(SFLA)is used for solving the NoC mapping problem based on the concept of adj ustment sequence.To further enhance searching capability,the proposed SFLA is improved with the adaptive chaos tactic and strategy of multi-neighborhood annealing. Experimental results shows that the improved SFLA has the advantage over the SFLA and annealing SFLA of better optimizing performance,thus reducing the communication power further.