Global Pattern Search for alpha-level Optimisation

The Global Pattern Search for alpha-level Optimisation (aGPS) is a global optimisation approach explicitly designed for efficient and user-friendly alpha-level optimisations. It can be used to improve the efficiency of fuzzy structural analyses. The efficiency of aGPS stems from its deterministic sample generation, which allows a reuse of many samples within the various alpha-level optimisations. Moreover, information gained within an alpha-level optimisation is used for all subsequent optimisations. It outperforms state-of-the-art algorithms. This means that it requires less model evaluations, and therefore, it has lower computing times. Here, the entire source code for its implementation in MATLAB is given. For further information, it is referred to "Huebler, C., & Hofmeister, B. (2021). Efficient and user-friendly alpha-level optimisation for application-orientated fuzzy structural analyses. Submitted to Engineering Structures."

BibTex: