Bridge Building with Grammatical Evolution

In this project Grammatical Evolution will be used to automatically design a Bridge with the aid of a Physics Engine such as the Open Dynamics Engine. Grammatical Evolution is a form of Genetic Programming that uses Grammars to dictate how solutions to a problem are constructed. It belongs to the broader field of Evolutionary Computation, which are a family of problem solving algorithms that draw inspiration from the natural process of biological Evolution. In these algorithms a population of candidate solutions to a problem are typically generated randomly. The quality of each candidate solution is measured, and this fitness metric is used during Selection to pick parents from which new child candidate solutions will be created. These progeny are usually created using recombination and mutation. A number of iterations (each iteration is referred to as a generation) of this selection and variety-generation process are conducted during which time the quality of the candidate solutions improve. The use of Evolutionary Computation for Automated Design represents a relatively new and fruitful application of these methods. In particular, Genetic Programming has enjoyed much success, for example, in the design of 3D Surfaces for Architects, Analog and Digital Circuits, and Antennae for NASA space missions. Some of the most notable applications in this area have resulted in the evolved solutions themselves being patentable. A Grammatical Evolution library in Java (GEVA) will be provided, which must be extended to allow interaction with the Physics engine.