3. Tactical planning problems
Tactical planning problems deal with optimally utilizing the given infrastructure by choosing services and
associated transportation modes, allocating their capacities to orders, and planning their itineraries and frequency.
Table 2 and Fig. 3 provide an overview of the literature discussed on tactical planning problems.
Deciding whether to send cargo direct or through a consolidation system entails a tradeoff influenced by system costs, operation times, network structure, and customer requirements. In the literature on tactical planning problems, mostly hub-and-spoke structures are regarded. Freight on hub-and-spoke networks is transported by a single service, or
a sequence of services where the loads are transferred from one service to the next at intermediate terminals. A service is characterized by its origin, destination, and intermediate terminals, its transportation mode, route, and its service capacity. Likewise, a mode is characterized by its loading capacity, speed, and price. Usually, these services and modes have fixed costs.
Following Fig. 3, there are two groups of models. The first group, Network Flow Planning (NFP), relates to the flow planning decisions addressing the movement of orders (commodities) throughout the network.
The second group, Service Network Design (SND), involves the service planning decisions including all decisions on choosing the transportation services and modes to move those commodities.
SND problems are furthermore partitioned into static and dynamic problems.
While in both groups one determines the frequency of the service, the capacity allocation, the equipment planning, and the routing and flow of commodities, in the former it is assumed that all problem aspects are static over the time horizon,
and in the latter, at least one feature (e.g. demand) varies over time.
Table 2 shows a growing body of literature in SND problems compared to NFP problems. This trend indicates that decision makers take the fixed cost of employing services into account and look for cost-efficient solutions.
In both NFP and SND problems, continuous variables are used to represent commodity flows throughout the network, but in SND problems, binary variables are included for selection of services.
The variables can be arc-based representing flow on an arc or path-based representing flow on a path (a series of arcs). SND problems are then modeled as Fixed-Cost Capacitated Multicommodity Network Design (CMND) problems.
Dynamic SND problems have a time dimension in the CMND formulation, making it a discrete multi-period model.
Therefore,the SND problem is mapped both in time and space, and each node in the new network represents a terminal at a time period. This space–time network has the potential to accommodate many real-life properties of SND problems. An example is waiting or transfer operations at a terminal, represented by arcs connecting different time periods with the same locations. Another example is different transportation modes, represented by additional arcs.
Such arcs can therefore accommodate different costs for terminal operations and modal costs. For an overview on dynamic SND problems, we refer to Wieberneit (2008).