A speciation-based bilevel niching method for multimodal truss design problems
Autor: | Xiaodong Li, Md. Jakirul Islam, Kalyanmoy Deb |
---|---|
Rok vydání: | 2021 |
Předmět: |
Mathematical optimization
Control and Optimization Optimization problem Computer science Applied Mathematics Truss Topology (electrical circuits) Network topology Computer Science Applications Computational Theory and Mathematics Robustness (computer science) Genetic algorithm Theory of computation Discrete Mathematics and Combinatorics Metaheuristic |
Zdroj: | Journal of Combinatorial Optimization. 44:172-206 |
ISSN: | 1573-2886 1382-6905 |
Popis: | Truss design is a well-known structural optimization problem that has important practical applications in various fields. Truss design problems are typically multimodal by nature, meaning that it offers multiple optimal solutions concerning the topology (combinatorial optimization problem) and/or sizes (continuous optimization problem) of the members, but they are evaluated to have similar or equally good objective function values. From a practical standpoint, it is desirable to find as many alternative designs as possible, rather than finding a single design, as often practiced. Several techniques based on classical and metaheuristic optimization methods have been developed for simultaneous optimization of topology and size of a truss. However, all these methods unable to find multiple topologies and their corresponding size solutions in a single run. A few metaheuristics incorporating niching techniques have been developed for finding multiple topologies for the truss design problem, but these studies ignored the fact that for each known topology, multiple design solutions in terms of size can be found. To address this issue, this paper proposes a bi-level truss formulation and subsequently a speciation-based bilevel niching method (BiL-NM) using such a formulation. The BiL-NM consists of a modified SPSO niching method which is robust to find multiple topologies and a canonical PSO for their corresponding size solutions. Extensive empirical studies are carried out to analyze the accuracy, robustness, and efficiency of the BiL-NM. The results confirm that the proposed BiL-NM is superior in all these three aspects over the state-of-the-art methods on several low to high-dimensional truss design problems. |
Databáze: | OpenAIRE |
Externí odkaz: |