Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!mitel!sce!bond From: bond@sce.carleton.ca (Greg Bond) Newsgroups: comp.ai Subject: ATMS implementation query Message-ID: <730@sce.carleton.ca> Date: 2 Dec 89 03:21:56 GMT Organization: Systems & Computer Eng. Dept., Carleton University, Ottawa, Canada Lines: 40 I'm in the midst of designing an ATMS and have run into difficulties surmising implementation details of typed conjunctive consumers from deKleer's third paper (Problem Solving with the ATMS). I would appreciate any help I can get (in the form of advice or pointers to papers, tech reports or books) because any solution to the problem that I can think of is rather kludgy - kludges do not belong in as elegant a mechanism as the ATMS. deKleer states that a typed consumer is executed only if the node receives a valid justification with another type. I understand how this concept applies to basic typed consumers but, assuming that they abide by the same principles of operation, I cannot see how this applies to typed conjunctive consumers. deKleer explains that a conjunctive consumer is installed as a temporary node with a basic consumer G. G invokes the conjunctive consumer F from the temporary node's justification informant when the temporary node is updated to hold in some environment. If G is typed to reflect the type of F then G will not be executed until the temporary node receives a valid justification with another type. Since it is a temporary node, it is invisible to the problem solver and should never be justified again. Thus, G will never be enqueued and, therefore, neither will F. If G is not typed then, when the temporary node's label is updated to hold in some environment, G will be executed and, therefore, so will F - regardless of F's type. The type checking procedure for typed conjunctive consumers should be checking the justification types of F's antecedents not the justification types of G's antecedents. This would mean using different type checking procedures for typed basic consumers and typed conjunctive consumers - but this is not the impression I get from reading deKleer's paper. Can anyone show me the error of my ways? ------------------------------------------------------------------ Greg Bond -----> bond@sce.carleton.ca Dept. of Systems and Computer Engineering, Carleton University Ottawa, ON, Canada K1S 5B6 Brought to you by Super Global Mega Corp .com