Multi-criteria path optimization based on a driver's preferences in route selection using a LLNFM combined with a GA

In this study, some different approaches are designed, implemented, and evaluated to solve a multi-criteria path optimization problem based on a driver’s preferences in multi-criteria route selection using a locally linear neuro-fuzzy model (LLNFM) with an incremental tree based learning algorithm combined with a multi-objective genetic algorithm. Using our designed neuro-fuzzy toolbox, the driver’s preferences in multi-criteria route selection such as the number of output classes and the routes in each output-class were received. Next, some approaches were proposed to classify each route based on the mentioned preferences chosen by the driver. Finally, the trained LLNFM-based models were utilized to classify Pareto-optimal paths obtained by applying a multi-objective genetic algorithm between an origin and a destination within the most dense-part of Tehran transportation network, i.e., districts 6,7,11, and 12, based on the driver's preferences in multi-criteria route selection. The results show the validity of the proposed approaches used to solve a multi-criteria path optimization problem based on a driver’s preferences in multi-criteria route selection.