DID YOU KNOW? DTIC has over 3.5 million final reports on DoD funded research, development, test, and evaluation activities available to our registered users. Click
HERE to register or log in.
Accession Number:
ADA183419
Title:
The Forbin Paper.
Descriptive Note:
Research rept.,
Corporate Author:
YALE UNIV NEW HAVEN CT DEPT OF COMPUTER SCIENCE
Report Date:
1987-07-01
Pagination or Media Count:
84.0
Abstract:
Planning is the process of formulating sequences of actions which, if carried out, can be expected to bring about certain situations. Search in planning involves simulating various sequences of actions to see whether or not they will bring about the desired situation. Since the space of possible sequences of actions is quite large in comparison with the set of sequences that actually bring about the desired situation, it is important to carefully guide the search. At each point in the planning process, the planner should have before it a representation that describes the planners intentions and their potential consequences to help in deciding how to refine those intentions. This is the essence of what has been called hierarchical planning. The result of each decision constitutes a commitment on the part of the planner, and as commitments are made during planning they form the basis for subsequent decisions. In most situations, it is impossible to guarantee that a decision, once made, is immune to retraction. In the sort of routine planning situations that we are concerned with, retracting previous commitments, or backtracking as it is commonly referred to, while it cant be avoided altogether, can be carefully controlled. One can, direct the decision making process so that, having performed a certain amount of search significantly less than that required for the entire problem, one can reliably commit to the consequences of a particular decision without fear of later retraction. This paper describes a planning architecture that supports a form of hierarchical planning well suited to applications involving deadlines, travel time, and resource considerations.
Distribution Statement:
APPROVED FOR PUBLIC RELEASE
