| This dissertation concerns the principles and techniques for scalable evolutionary computation to achieve better solutions for larger problems with more computational resources. It suggests that many of the limitations of existent evolutionary algorithms, such as premature convergence, stagnation, loss of diversity, lack of reliability and efficiency, are derived from the fundamental convergent evolution model, the oversimplified "survival of the fittest" Darwinian evolution model. Within this model, the higher the fitness the population achieves, the more the search capability is lost. This is also the case for many other conventional search techniques.; The main result of this dissertation is the introduction of a novel sustainable evolution model, the Hierarchical Fair Competition (HFC) model, and corresponding five sustainable evolutionary algorithms (EA) for evolutionary search. By maintaining individuals in hierarchically organized fitness levels and keeping evolution going at all fitness levels, HFC transforms the conventional convergent evolutionary computation model into a sustainable search framework by ensuring a continuous supply and incorporation of low-level building blocks and by culturing and maintaining building blocks of intermediate levels with its assembly-line structure. By reducing the selection pressure within each fitness level while maintaining the global selection pressure to help ensure exploitation of good building blocks found, HFC provides a good solution to the explore vs. exploitation dilemma, which implies its wide applications in other search, optimization, and machine learning problems and algorithms.; The second theme of this dissertation is an examination of the fundamental principles and related techniques for achieving scalable evolutionary synthesis. It first presents a survey of related research on principles for handling complexity in artificially designed and naturally evolved systems, including modularity, reuse, development, and context evolution. Limitations of current genetic programming based evolutionary synthesis paradigm are discussed and future research directions are outlined. Within this context, this dissertation investigates two critical issues in topologically open-ended evolutionary synthesis, using bond-graph-based dynamic system synthesis as benchmark problems. For the issue of balanced topology and parameter search in evolutionary synthesis, an effective technique named Structure Fitness Sharing (SFS) is proposed to maintain topology search capability. For the representation issue in evolutionary synthesis, or more specifically the function set design problem of genetic programming, two modular set approaches are proposed to investigate the relationship between representation, evolvability, and scalability. |
