现代微波/射频系统通常希望微波器件能够多功能工作，展现良好的结构适应性，并具有尽可能好的性能。相应地，就需要无源器件有新的结构和新的设计技术。多目标优化技术在此类器件的设计中有着极大的潜力，因为采用多目标优化可以同时实现多个指标趋优。但是，设计中要求多目标优化算法能够保证解的多样性和优化搜索的高效率。本项目旨在研究如何采用离散化碎片结构，利用多目标优化搜索来寻找最佳拓扑结构、设计逼近指标极限的无源器件。为此，我们首先要研究有耗网络S参数的约束以及实际电路的指标极限，接着研究如何兼顾优化搜索要求，采用离散化碎片结构来描述具有任意拓扑结构的微波无源器件，还要研究相应的多目标优化策略和算法，尽可能高效地找出一个在给定的设计空间中可以逼近指标极限的电路结构。通过这些理论和方法探索，为结构适应性好、各项性能趋优的小型化高性能微波无源器的件设计，提供一种全新的、高效的设计技术。

In modern microwave/RF systems, passive devices are generally desired to have multiple functions, satisfactory structure versatility, and overall optimal performance. Accordingly, novel structures and novel design techniques for passive devices are always required. Multi-objective optimization techniques show great potentials in design of such microwave/RF devices because the design is generally a multi-objective optimization. Multi-objective optimization must provide desired diversity of solution of a multi-objective optimization problem, and high efficiency of optimization searching in decision space. In this project, we focus on the design of passive microwave/RF devices with fragmented structure and arbitrary topology, and performance approaching to the extremes for a circuit with practical losses. We will first investigate the limits of scattering parameters of a lossy network and the extreme of characteristics or specifications of the corresponding circuit. Then by using a proper multi-objective optimization searching technique, we will investigate how to find a circuit structure, within a given design space of specific dimensions and shape, that may provide or approach to the extreme characteristics. With the proposed theory and optimization technique, we hope we can find an efficient and useful technique for design of passive microwave/RF circuits of the highest performance and the best structure.