The response time of an algorithm is the time which is needed until a newly calculated solution can be applied. Algorithms for dynamic planning must have fast response times for two reasons. First, a solution calculated for a dynamic problem can only be applied if the input data have not changed significantly during the planning process. Second, the longer it takes to calculate a new solution the higher is the probability that dispatchers concurrently change the current solution.
In many cases dispatchers have to decide about load acceptance or rejection while the shipper is on the phone making the request. Dynamic planning systems should be capable of supporting dispatchers in the load acceptance decision while the shipper is on the phone. That is, within a couple of seconds they should give support in deciding whether a transportation request can be served feasibly and efficiently.
The main advantage in dynamic planning is that there is usually plenty of time for optimisation. Ten minutes spent to find a solution for a small problem may seem like a long time in the static case. In the dynamic case, however, one has all day so the time might as well be used. A dynamic planning system can be used to successively improve the current solutions. Obviously, fast response times can not be achieved if new solutions are calculated from scratch every time the dynamic planning method is invoked. Therefore, a dynamic planning system should have a restart capability, i.e. the planning system should be able to continue from the current solution. Furthermore, dynamic planning methods require efficient information update mechanisms in order to efficiently consider modified input data.