Method for automatic generation of optimal space frame
Space frames consist of combination of one or more basic modules to function as a decorative feature or a building element in architectural design. There are certain aesthetic criteria to meet on the design of a basic space frame module; space frame module, geometric regularity, and dimension regularity for all the angles and lengths between edges. Designers exercise their own creativity to develop space frame modules base on these criteria. The present invention relates to a hybrid algorithm based on evolutionary algorithm for graph encoding scheme (EAGES) and genetic algorithm (GA) to evolve the design automatically in relative small number of generations. The hybrid algorithm is a tool to architects and designers to rapidly produce aesthetically pleasing designs with the resources available to them.
1. A method of designing space frames with at least four nodes used as an architectural, decorative feature, comprising the steps of:
randomly initializing a population of graphs;
encoding each graph in the population into an adjacent matrix;
operating on said graphs encoded in adjacent matrices using an evolution algorithm for graph encoding scheme (EAGES) which is specifically designed for evolution of adjacent matrices;
obtaining architecturally optimized topological space frame module designs;
initializing linear chromosomes on three-dimensional coordinates of said optimized space frame module designs;
operating on said linear chromosomes using a genetic algorithm (GA); and
obtaining a space frame module having optimal, topology, geometry, and dimensions.
2. The method according to claim 1 , wherein said evolution algorithm for graph encoding scheme (EAGES) comprises utilizing a selection operator, crossover operator, followed by a mutation operator, followed by a termination operator, wherein all operators operate on the adjacent matrices.
3. The method according to claim 2 , wherein said selection operator is a roulette wheel selection.
4. The method according to claim 2 wherein said crossover operator is a random crossover in an evolutionary algorithm for graph encoding scheme (EAGES) which operates on the adjacency matrices.
5. The method according to claim 2 , wherein said mutation operator is selected from the group consisting of a number-of-edge operator, a node-label operator, and a swap-node operator, wherein all operators operate on the adjacent matrices.
6. The method according to claim 2 , wherein said termination operator is conditioned on the target fitness occurring after 100 generations or when 1 is reached.
7. The method according to claim 1 , wherein said genetic algorithm incorporates using a selection operator, crossover operator, mutation operator, and termination operator.
8. The method according to claim 7 , wherein said selection operator is a roulette wheel selection.
9. The method according to claim 7 , wherein said crossover operator is a uniform crossover.
10. The method according to claim 7 , wherein said mutation operator is uniform mutation.
11. The method according to claim 7 , wherein the termination operator is conditioned on the target fitness occurring to a maximum of 1000 generations or when the target fitness is 1.
12. The method according to claim 1 , further comprising the steps of:
prior randomly initializing a population of graphs;
hybridizing said evolution algorithm for graph encoding scheme (EAGES) and said genetic algorithm (GA); and
storing said hybridized algorithms on the memory of a computer system.