SOft Computing applications for Complex EnviRonments

Multimodal optimization:

an effective framework for model calibration
Co-authored by Manuel Chica, José Barranquero, Óscar Cordón, Sergio Damas and Tomasz Kajdanowicz Corresponding author contact:

Automated calibration is an important stage when validating non-linear dynamic systems. This is not a blind process as the modeler must control and analyze the calibration results and parameters values in an iterative way. In many non-linear models such as system dynamics it is usual to find sets of configuration parameters that may obtain the same model fitting. In these cases the modeler needs to understand the results implications and run sensitivity analysis to check the model validity. We propose a framework based on niching genetic algorithms with evaluation, visualization, and filtering processes to assist the modeler in the validation phase by returning a set of alternative calibration solutions. The core component of the framework is the niching process, able to reach various optima in multimodal optimization problems by keeping the necessary diversity. We illustrate the way to apply the proposed framework in two case studies, different both in application field and modeling methodology. The first case study is a biological growth model. The second one is a managerial model for improving the brand image of a TV show. The results of this work demonstrate the benefits of the proposed framework with respect to standard automated approaches for model calibration.

Link to the paper.

Source code in C++ and Matlab evaluation wrapper Download Code.

This Work is supported by Spanish Ministerio de Economía y Competitividad under the NEWSOCO project (ref. TIN2015-67661), including European Regional Development Funds (ERDF), and by the European Commission under the 7th Framework Programme, Coordination and Support Action, Grant Agreement Number 316097, Engine project.