Medevac Domain Object Model

  
Figure 3.1: Medevac Model OMT Diagram

The medical evacuation domain was modeled by selecting a subset of the DITOPS core concepts and specializing them for this problem. This chapter will present the object model developed for this domain. Standard DITOPS concepts are referred to but not documented here. The reader is referred to the previous chapter for a description of the standard DITOPS class library. The OMT diagram of figure 3.1 describes the object model of the medical evacuation domain. In this diagram, DITOPS core classes have their names in uppercase, while the medical evacuation domain classes (in essence, classes designed and implemented for this prototype) have their names in lowercase.

The general structuring of the problem is as follows:

• The system expects five different types of input: (1) descriptions of patients to be transported, (2) a set of scheduled missions flown by a set of aircraft (possibly with some capacity already allocated), (3) a set of hospitals with specified medical specialties, (4) a set of airfields, with associated ASFs; finally, (5) disruptive events which modify the systems representation of the world and have to be reacted to.
• The system will perform both generative and reactive scheduling. In the generative mode, routes (itineraries) are created for each unscheduled patient. In the reactive mode, disruptions in the existing schedule are being reacted to by repairing the schedule.

The general modeling approach taken can be described as follows:

• Patients are modeled as demands, incoming orders to the system.
• Aircraft are modeled as batch resources (multiple-capacity resources with time-synchronized execution). They are also ``movable'', meaning that they keep track of their geographical position.
• Airfields, Medical Treatment Facilities and Air Staging Facilities are modeled as aggregate resources (multiple-capacity resources with no requirement for time-synchronized execution).
• Missions, mission legs and medical treatment operations are modeled as transport operations. These operations potentially change the position of their executing resource (such as an aircraft).
• Constraints between mission legs (both from the missions' standpoint and from the patients' standpoint) were modeled as types of before temporal relations.

It should be noted that contrary to the normal approach of scheduling operations formed as a response to receiving demands (orders), the scheduling process in this prototype has been organized somewhat differently: operations are already scheduled (missions, mission legs), and the role of the scheduler is to find suitable combinations of operations (routes) to satisfy as many demands as possible as well as possible. This can be seen as the mission legs in essence being resources, since they impose allocation constraints on the ``real'' resources, the aircraft.

• Object Model Description