通过将光电 MCM的功耗约束转化为光互连数最少的问题 ,建立了芯片级划分模型 ,解决了用遗传算法处理结构化设计的芯片级划分的问题 ,提出了基于功耗最小的光电 MCM划分优化算法 .实际的设计结果表明 ,该算法具有更强的寻优能力 ,比以往的算法更适合光电 MCM的划分 ,可使系统功耗降低 50 %以上 . By translating the power constraints of the optoelectronic MCM into those with the lowest number of optical interconnects, a chip-level partitioning model is established, which solves the problem of using the genetic algorithm to process the chip-level partitioning of the structured design. The optoelectronic MCM The actual design results show that the proposed algorithm has better searching ability and is more suitable for the partition of optoelectronic MCM than previous algorithms, which can reduce the system power consumption by more than 50%.