4.3 Segmenting the Execution Trace

The purpose of segmenting the execution trace is to identify sequences of actions that are logically grouped together. Currently, two kinds of groupings are identified: goal-groupings and time-groupings. In the future, location-groupings may also be considered. The segmentation process takes a cleaned execution trace as input and returns a list of segments-of-execution-trace/goals pairs as output.

Using the active goal information recorded in the execution trace makes identifying goal-groupings trivial: conglomerate all actions associated with identical goals. Identifying time-groupings is slightly more complicated. A consecutive subsequence of the (chronologically-sorted) trace is considered a time-grouping if the joint effects of the executed actions achieves the conjunction of the goals associated with them.

  

Figure 6: Schematic of a time-grouping segment for G1 and G2.

These two types of segments are represented schematically in Figure 6. The axis represents a portion of an actor's activity time-line with markers indicating the execution of actions A1 through A9 (more actions may precede or follow these actions). If the actor, individually or in cooperation with other actors, accomplished goals G1 and G2 during the intervals marked on the time-line, there would be three segments identified from this activity time-line. There would be a goal-grouping segment for G1 of (A2, A3, A7, A8), a goal-grouping segment for G2 of (A4, A5, A6) and a time-grouping segment for G1 and G2 of (A2, A3, A4, A5, A6, A7, A8). This last segment is the simplest type of segment which is a time-grouping segment that is not also a goal-grouping segment.