Abstract

A novel sequential sampling algorithm is described and tested using tensor decomposition as surrogate model through the definition of an objetive function that balances exploitation and exploration criteria.