Transportation Planning in Dynamic Environments

Summary Transportation planning in realistic scenarios has to deal with a dynamic uncertain environment. The actual state of all information relevant to planning can be estimated and be the base of a static planning environment which is often the base for classic planning algorithms. If the environment differs from the one used for planning, either by a wrong estimation or due to the dynamics in the environment the actual plan has to be adapted. Dynamic transportation planning is currently an active research area; nevertheless it is not a new problem in the literature. A good overview of the development of this field can be found in [1]. Different technologies like agents or meta-heuristics and their possible integration are in discussion [2]. Even if a lot of research has been done on the question how these problems can be solved, the dynamics has been rarely tried to measure. Commonly, dynamic transportation planning is characterized by the occurrence of disturbing events that invalidate the current plan. What kinds of events occur depends on the planning problem that was assumed. In the current state of research different technologies claim to be fruitful in more dynamic environments. Larsen [3] points out the importance of measuring dynamics for effi cient and sufficient performance comparison. Therefore he defines the degree of dynamism (DoD).Moreover, from an engineering perspective it would be very useful to have metrics and measurements at hand that can help to choose one technology for a given problem. The vision is that based on the characteristics of the given application one could decide which is the most appropriate technology. Therefore, different questions have to be answered, like how can dynamics be measured to what DoD classical approaches are competitive, and how different technologies perform on a varying DoD. In this article we tackle the first question. Therefore, we discuss here different existing approaches to measure dynamics in transportation planning. Based on the well-known multiple depot vehicle routing problem with time windows (MDVRPTW) problem, we present an extended version of the measurement of the DoD. In section 3, we present empirical results and Finally we discuss our Findings.