Belief and fuzzy theories for driving behavior assessment in case of accident scenarios

The estimation of the overspeed risk before the accident is among the main goals of this paper. The proposed method uses the Energy Equivalent Speed (EES) to assess the severity of an eventual accident. However, the driver behavior evaluation should take into account the parameters related to the Driver, the Vehicle and the Environment (DVE) system. For this purpose, this paper considers a two-level strategy to predict the global risk of an event using the Dempster-Shafer Theory (DST) and the Fuzzy Theory (FT). This paper presents two methods to develop the Expert Model-based Basic Probability Assignment (EM based BPA), which is the most important task in the DST. The first one is based on the accident statistics and the second method deals with the relationship between the Fuzzy and Belief measurements. The experimental data is collected by one driver using our test vehicle and a Micro-intelligent Black Box (Micro-iBB) to collect the driving data. The sensitivity of the developed models is analysed. Our main evaluation concerns the Usage Based Insurance (UBI) applications based on the driving behavior. So, the obtained masses over the defined referential subsets in the DST are used as a score to compute the driver’s insurance premium.     

A hybrid Genetic Algorithm approach to minimize the total joint cost of a single-vendor multi-customer integrated scheduling problem

This paper addresses the scheduling of supply chains with interrelated factories consisting of a single vendor and multiple customers. In this research, one transporter is available to deliver jobs from vendor to customers, and the jobs can be processed by batch. The problem studied in this paper focuses on a real-case scheduling problem of a multi-location hospital supplied with a central pharmacy. The objective of this work is to minimize the total cost, while satisfying the customer’s due dates constraints. A mathematical formulation of the problem is given as a Mixed Integer Programming model. Then, a Branch-and-Bound algorithm is proposed as an exact method for solving this problem, a greedy local search is developed as a heuristic approach, and a hybrid Genetic Algorithm is presented as a meta-heuristic. Computation experiments are conducted to highlight the performance of the proposed methods.