Coevolution in Complex Adaptive Systems

这一小笔探索性研究拨款将支持一项调查，该调查将探索到混沌边缘的共同进化的基础和含义，该调查基于分子或商品作为符号串的全新模型，这些分子或商品相互作用，根据“随机语法”产生新的符号串。首席研究员寻求对复杂适应系统如何协作、竞争和共同进化的新见解。特别是，研究人员希望在一个经济体中实现足够多样化的商品和服务的基础上，找到新的经济起飞模式。类似的概念也将适用于生态系统结构。更广泛地说，人们希望对随机语法的研究将产生对远离平衡系统的自组织行为的洞察。在生物学和其他经济和文化系统中的基本现象中，复杂适应系统的共同进化。在过去的几年里，少数物理学家、生物学家、数学家和经济学家在不同领域开展的工作，已经开始提出一种广泛的新概念框架和这种共同进化的候选普遍特征：复杂的共同进化系统可能会进化到混沌的边缘。这一新理论建立在这样一种认识之上，即这种共同进化的系统，无论是共生体还是相互关联的公司，都共同构建了它们共同居住的世界。这样的共同进化系统可以在三种机制中发挥作用。首先是一个混乱的制度，每个适应性代理人都在努力改善多峰的“适应景观”，由于其他代理人的移动，这些景观变形得如此之快，以至于没有一个代理人停下来。这种制度在进化生物学中被称为军备竞赛。当所有代理都爬到相互一致的局部适应度峰值时，就会出现有序的制度。适应性机会停止了。第三个政权是秩序和混乱之间的分界线。出现了两个新的事实：1)平均适合度，或“回报”在混乱和有序的制度下都很低，在混沌的边缘是最优的。2)一种“看不见的手”可以起到这样的作用，即每个主体调整自己的景观的结构，并将其与其他景观的耦合转化为自己的优势，通过这样做，整个系统共同进化到秩序和混沌之间的相变。如果大体上属实，其影响是广泛的。我们必须考虑到，适应性因素相互改变了我们彼此之间的游戏，无论是在进化中，还是在经济中，或许在其他地方，都是在秩序和无序之间的一种特有的平衡状态。这种稳定状态的特征包括由微小扰动引发的雪崩变化的幂分布。在进化记录中，灭绝事件的大小分布就是已知的。类似的雪崩可能会发生在技术进化过程中。

This Small Grant for Exploratory Research will support an investigation that will pursue the underpinnings and implications of the coevolution to the edge of chaos, based on entirely novel models of molecules or goods as symbol strings which act on one another to produce new symbol strings according to a "random grammar". The principal investigator seeks new insights into how complex adaptive systems collaborate, compete, and coevolve. In particular, the investigator expects to find new models of economic take-off based on achieving a sufficient diversity of goods and services in an economy. Similar concepts will apply to ecosystem structure. Most broadly, it is hoped that investigation of random grammars will yield insight into the self organized behaviors of far from equilibrium systems. Among the fundamental phenomena in biology and beyond, in economic and cultural systems, is the coevolution of complex adaptive systems. In the past few years, work carried out on diverse fronts by a small number of physicists, biologists, mathematicians, and economists, has begun to suggest a broad new conceptual framework and candidate universal features of such coevolution: Complex coevolving systems may evolve the the edge of chaos. The new theory is built upon the recognition that such coevolving systems, whether symbionts or interlocked firms, mutually build the world they jointly inhabit. Such coevolutionary systems can behave in three regimes. The first is a chaotic regime, where each adaptive agent struggles to improve on multipeaked "fitness landscapes" which deform so rapidly due to moves of other agents that no agent comes to rest. This regime is called the Arms Race in evolutionary biology. The ordered regime occurs when all agents climb to local fitness peaks that are mutually consistent. Adaptive chance ceases. The third regime is the boundary between order and chaos. Two new facts have emerged: 1) Mean fitness, or "payoff" is low in both the chaotic regime and in the ordered regime and is optimal at the edge of chaos. 2) A kind of "invisible hand" can act such that each agent tunes the structure of its own landscape and couplings to other landscapes to its own advantage, and by doing so the entire system coevolves to the phase transition between order and chaos. If generally true, the implications are broad. We must consider that adaptive agents mutually alter the games we play with one another, in evolution, in economies, and perhaps elsewhere, to a characteristic poised state between order and disorder. Signatures of this poised state include a power law distribution of avalanches of change initiated by small perturbations. Just such a distribution of sizes of extinction events in the evolutionary Record are known. Similar avalanches may occur in technological evolution.