6.6.5 Design tradeoffs

As described in Section 2, a balance must be struck between the capabilities provided and resources used. The tradeoffs are different in every application and are usually a critical aspect of the design of an execution assistant. In the SUO domain, terrain reasoning is a key factor in this tradeoff. Using fine-grained terrain data to analyze progress or project future failures can overload computational resources. Therefore, the EA uses coarse terrain reasoning, but our design allows higher-fidelity terrain reasoners to respond to a defined set of terrain analysis requests. This feature allows the system to adjust its analysis to the tempo of operations.

Other key features to consider when making tradeoffs between reactivity and capabilities are the amount of processing done by the mission-monitoring methods, the report-monitoring Acts, and any specialized reasoners (such as terrain reasonsers) invoked by the methods or Acts. The user can adjust the frequency of monitoring at any time by customizing parameter settings. Currently, the SUO EA is not computationally overburdened while analyzing every report in full, but adding more computationally expensive projections or alerts in the future could cause reconsideration of this design decision. Finally, the amount of filtering of incoming reports done by the Watchman agent affects this balance.