TIE Description


In the general scenario, a ForMAT user receives the description of a new mission. In order to build a plan for the new mission, the user queries ForMAT's database of plans. While browsing ForMAT's plans, the user refines the mission statement in terms of specific objectives, i.e., goals to be achieved. ForMAT sends messages in real time to PRODIGY/Analogy capturing the history of the user actions. The user states the mission goals using a goal language that we introduced in ForMAT. Goals are of different types and can take a variety of arguments. An example of such goal may be ``send ground-unit to location.''

When the mission goals are entered by the user, ForMAT sends this information to PRODIGY/Analogy. PRODIGY/Analogy uses this information and processes it according to its domain model. Using the mission's goals, PRODIGY/Analogy retrieves, from its own database of plans (obtained from the ForMAT's database), similar past solved plans. It is further capable of identifying the needed modifications to the past plans as a function of the new and past missions' rationale. This information is sent again in real time to ForMAT as suggestions to the user on the plans and corresponding modifications that the user should do to build a plan for the new mission.

ForMAT supports the following capabilities:

  1. User receives the text description of a new mission;
  2. User queries database of past plans;
  3. User browses past plans;
  4. User refines query incrementally;
  5. User uses objective representation to specify the goals of the new mission;
  6. Plans and their Force Modules (FMs) are stored, indexed by goals and scenario information.
In PRODIGY (and PRODIGY/Analogy), the following features are used:
  1. Objects (such as units, locations, FMs) are organized in a class ontology;
  2. Plan actions are modeled and represented as plan operators;
  3. Planner can generate plans at a high-level of abstraction;
  4. Plans are stored, indexed by goals and scenario;
  5. Plans are retrieved based on goal similarity;
  6. Rationale is used to propose appropriate modifications.
The integrated ForMAT/PRODIGY systems provides mixed-initiative plan retrieval and modification through the following features:
  1. ForMAT and PRODIGY exchange messages in real time;
  2. ForMAT sends user's actions to PRODIGY;
  3. PRODIGY can follow user's plan browsing history;
  4. ForMAT's user saves goals: PRODIGY plans at the force module level;
  5. PRODIGY retrieves plans with similar objectives (goals);
  6. PRODIGY sends suggested plan modifications driven by plan rationale;
  7. ForMAT's user processes received mission statement.

Correspondence: mcox+@cs.cmu.edu

Last Edited: August 16, 1996