A Bayesian algorithm selection framework for black box optimization problems is proposed. A set of benchmark problems is used for training. The performance of a set of algorithms on the problems is recorded. In the beginning, an algorithm is randomly selected to run on the given unknown problem. A Bayesian approach is used to measure the similarity between problems. The most similar problem to the given problem is selected. Then the best algorithm for solving it is suggested for the second run. The process repeats until n algorithms have been run. The best solution out of n runs is recorded. We have experimentally evaluated the property and performance of the framework. Conclusions are (1) it can identify the most similar problem efficiently; (2) it benefits from a restart mechanism. It performs better when more knowledge is learned. Thus it is a good algorithm selection framework.
|Title of host publication||Simulated Evolution and Learning : 11th International Conference, SEAL 2017, Proceedings|
|Editors||Yuhui SHI, Kay Chen TAN, Mengjie ZHANG, Ke TANG, Xiaodong LI, Qingfu ZHANG, Ying TAN, Martin MIDDENDORF, Yaochu JIN|
|Publisher||Springer-Verlag GmbH and Co. KG|
|Number of pages||12|
|Publication status||Published - 2017|
|Event||11th International Conference on Simulated Evolution and Learning, SEAL 2017 - Shenzhen, China|
Duration: 10 Nov 2017 → 13 Nov 2017
|Name||Lecture Notes in Computer Science|
|Conference||11th International Conference on Simulated Evolution and Learning, SEAL 2017|
|Period||10/11/17 → 13/11/17|
Bibliographical noteFunding Information:
Acknowledgement. The work described in this paper was supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China [Project No. CityU 125313]. Yaodong He acknowledges the Institutional Postgraduate Studentship from City University of Hong Kong. Yang Lou acknowledges the Institutional Postgraduate Studentship and the Institutional Research Tuition Grant from City University of Hong Kong.
© Springer International Publishing AG 2017.
- Algorithm selection
- Bayesian approach
- Evolutionary algorithm
- Monte Carlo method
- Optimization problems