"Many fields of science and engineering have come against the problem of complex design. This occurs when the details become so numerous that computer power alone cannot make progress. For neural networks (aka deep learning), this is known as the NP-completeness problem whereby the amount of computation needed to compute network weights grows in a non-polynomial (i.e. exponential) way as the number of weights increases. For systems biology, it happens when there are too many cell-model parameters to allow for successful estimation. For many engineering fields, it happens when the designs become so complex that integrated testing cannot eliminate faults. Nature solved this complex-design problem using evolution. But exactly how it solved it has been a mystery. Both laboratory experiments and computer-simulation attempts eventually stopped evolving. Something more than Darwin's ideas of heredity, variation, and selection was needed. The solution is that there is a fourth element to evolution and that is ecological change. What happens is that a new variation that is selected can change the ecology and the new ecology can create new opportunities for even more new variations to be selected. By this endless cycle, complexity can grow automatically. This book uses the physics of resource flow to describe in detail how that happens and that knowledge can be used to help overcome the problem of complex design. This book is a monograph that can be used as a textbook on how physics plays a strong role in evolution. The key starting ideas are the flow of energy-and-matter resources, organisms making copies of themselves, and ecological change. From these ideas, quantitative models are developed for many evolutionary processes. Such models include that for selection, multicellularity, coevolution, sexual reproduction, and the Serengeti Rules. Numerous examples of biological phenomena are shown to be in conceptual agreement with the models. In the end, the physics shows how complex design can arise automatically"--