Compact real-valued teaching-learning based optimization with the applications to neural network training

作者：

Highlights：

• A new compact teaching-learning based optimization method is proposed.

• 32 well-known benchmark tests are conducted for algorithm evaluation.

• The algorithm maintains high performance and reduce memory requirement.

• The algorithm is used to train two typical function neural networks.

摘要

•A new compact teaching-learning based optimization method is proposed.•32 well-known benchmark tests are conducted for algorithm evaluation.•The algorithm maintains high performance and reduce memory requirement.•The algorithm is used to train two typical function neural networks.

论文关键词：

论文评审过程：Received 9 January 2018, Revised 1 June 2018, Accepted 2 June 2018, Available online 4 June 2018, Version of Record 10 September 2018.

论文官网地址：https://doi.org/10.1016/j.knosys.2018.06.004