Routing strategy in a distribution network: fixed route with driver learning versus variable daily optimized route

This paper faces with a critical short term decision: the construction of routes for daily deliveries in a distribution network. Typically, for each customer it is possible to have, in function of its requirements and demand pattern, daily deliveries, regular non-daily deliveries (i.e. once a week), and irregular deliveries. In these circumstances if daily routing optimization strategy is applied, it is possible that each route keep changing. This implies that each driver daily visits different customers, facing with different delivery areas, different acceptance procedures, different load-unload facilities, different operators to work with. We define the time spent in this kind of activities as “service time” whereas the time for the transportation from a generic point A to a generic point B is defined “travel time”. The service time could be reduced if each driver keeps familiarity with the served territories, for example by making fixed routes. This reduction depend on the level of learning-knowledge that each driver has developed during the past visits to the considered customers, whose function is explicable using learning curves. The authors investigate the possibility to apply a fixed routing strategy instead of a daily routing optimization strategy, analyzing the benefits derived from the driver familiarity according to the variability of the customer demand in quantity and frequency and in function of the ratio between service time and travel time. In this way the paper aims to provide an effective and flexible decision making tool in order to identify, for a considered distribution network, the best routing strategy (fixed/daily optimized) to apply considering the level of learning of the route possessed by each driver.