Automatic generation of algorithms for robust optimisation problems using Grammar-Guided Genetic Programming

We develop algorithms capable of tackling robust black-box optimisation problems, where the number model runs is limited. When a desired solution cannot be implemented exactly aim to find one, worst case in an uncertainty neighbourhood around still performs well. To investigate improved methods we employ automatic generation approach: Grammar-Guided Genetic Programming. algorithmic building blocks Particle Swarm Optimisation framework, define rules for constructing heuristics from these components, and evolve populations search problems. Our combine elements existing techniques new features, resulting investigation novel heuristic space. obtain which improve upon current state art. also analyse component level breakdowns developed against their performance, identify high-performing components