Evolving controllers and controlling evolution

生物系统，特别是细胞相互作用网络，显示出复杂的动力学和广泛保守的结构特征，如模块性和鲁棒性。在各种工业应用中的工程系统和网络中发现了许多相同的动力学和结构特征4然而，至关重要的是，尽管最终结果在某些方面看起来非常相似，但工程系统和自然系统实现成功设计的过程非常不同-在工程中，模块化和鲁棒性在一个给定的系统中的结果，从使用正式的设计过程（系统和控制工程）;在生物学，从进化的过程。在这个项目中，我们建议利用协同效应，并带来控制理论和进化理论的领域之间的工具和思想的交叉受精，以解决以下具体问题：（1）什么进化压力引起了生物系统的鲁棒性？生物系统的进化主要是为了变得健壮，还是健壮是作为其他更重要特征的副产品而出现的？(2)基于模块的设计是合成生物系统工程的最佳方法吗？生物系统和突变算子的关键特征是否被进化用作设计工具？(3)有可能控制进化吗？我们能否将控制工程工具应用于进化过程，以生成具有所需特性的合成生物系统？(4)有没有进化原理可以用来设计更好的工程系统？我们能否将联合收割机进化仿真与数学优化结合起来，同时进化大型网络控制系统的结构和参数值，以最佳地满足相互冲突的设计标准？我们相信，即使是对这些问题的部分答案也将导致系统生物学，合成生物学以及系统和控制工程领域的变革性研究突破。

Biological systems, in particular cellular interaction networks, display complex dynamics and widely-conserved structural features such as modularity and robustness. Many of the same dynamical and structural features are found in engineered systems and networks across a diverse range of industrial applications4 Crucially, however, although the end-results may in certain respects look very similar, the processes by which engineered and natural systems arrive at successful designs are very different - in engineering, modularity and robustness in a given system result from the use of formalised design processes (systems and control engineering); in biology, from the process of evolution. In this project, we propose to exploit synergies and bring about cross-fertilisation of tools and ideas between the fields of control theory and evolutionary theory to address the following specific questions: (1) What evolutionary pressures have given rise to robustness in biological systems? Have biological systems evolved primarily to be robust, or has robustness arisen as a by-product of other more important characteristics? (2) Is module-based design an optimal approach to the engineering of synthetic biological systems or are there alternatives? Are there key features of biological systems and mutational operators that evolution exploits as a design tool? (3) Is it possible to control evolution? Can we apply control engineering tools to the evolutionary process to generate synthetic biological systems with desired characteristics? (4) Are there evolutionary principles that can be used for designing better engineering systems? Can we combine evolutionary simulation with mathematical optimization to simultaneously evolve both the structure and parameter values of large networked control systems to optimally satisfy conflicting design criteria? We believe that even partial answers to these questions will lead to transformative research breakthroughs across the fields of systems biology, synthetic biology and systems and control engineering.