Real Vehicle Routing and Dispatching with Dynamic Fuzzy Travel Times

The paper presents a model of the real world vehicle routing and dispatching problem. In real world, vehicle travel speeds are varying with period of time and can not be expressed as exact values. So the time-dependent and fuzzy travel speeds are introduced into the model. A dispatching period is divided into some time slices and each time slice is designated a triangular fuzzy speed. The travel time is calculated by fuzzy value integral to insure FIFO property. The method of comparing two triangular fuzzy numbers is applied to check whether or not customers´ time windows are satisfied. A hybrid intelligent approach combining genetic algorithm and ant colony algorithm is proposed for solving the dispatching model. A chromosome of the genetic algorithm is regarded as a ant of ant colony algorithm, and its fitness is regarded as the pheromone that remains on way that it passes. And chromosomes are constructed by searching path method of ant colony algorithm. The approach is viable and valid by computing a real instance.