Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Evolutionary quantum algorithms for structural design
Two genetic quantum-based algorithms are proposed for large scale problem analysis, and are compared with elite GA and a hybrid algorithm of GA and neural network (NN) in terms of computational efficiency with equal or better performance. Genetic quantum algorithms (GQA) similar to genetic algorithms (GA) maintain a population of individuals but each individual is composed of probabilistic quantum bits for preserving diversity, i.e. each individual of length m is equivalent to 2m states. Also instead of crossover or mutation, GQA use quantum gates (QG) to update individuals and to guide the evolutionary process. The role of NN is to replace the time consuming state of finite element analysis, but the neuro-approximation introduces error in fitness prediction as well. In comparison, statistical analysis reveals that GQA is not only relatively simpler, but it also decreases optimization time while finding better solutions. Furthermore, the usage of neural networks helps further reduce finite element evaluations without significantly compromising quality of solutions.
Evolutionary quantum algorithms for structural design
Two genetic quantum-based algorithms are proposed for large scale problem analysis, and are compared with elite GA and a hybrid algorithm of GA and neural network (NN) in terms of computational efficiency with equal or better performance. Genetic quantum algorithms (GQA) similar to genetic algorithms (GA) maintain a population of individuals but each individual is composed of probabilistic quantum bits for preserving diversity, i.e. each individual of length m is equivalent to 2m states. Also instead of crossover or mutation, GQA use quantum gates (QG) to update individuals and to guide the evolutionary process. The role of NN is to replace the time consuming state of finite element analysis, but the neuro-approximation introduces error in fitness prediction as well. In comparison, statistical analysis reveals that GQA is not only relatively simpler, but it also decreases optimization time while finding better solutions. Furthermore, the usage of neural networks helps further reduce finite element evaluations without significantly compromising quality of solutions.
Evolutionary quantum algorithms for structural design
Akbarzadeh-T, M.R. (Autor:in) / Khorsand, A.R. (Autor:in)
2005
6 Seiten, 16 Quellen
Aufsatz (Konferenz)
Englisch
Parallel evolutionary algorithms for design optimization
| British Library Conference Proceedings | 2003
Design of FRP laminates using evolutionary algorithms
| British Library Conference Proceedings | 2007
Evolutionary computation in structural design
| Tema Archiv | 2000
Optimization of retaining wall design using evolutionary algorithms
| British Library Online Contents | 2017