Jam agents in a Nutshell Version 65 76i 1
Download 391.89 Kb.
|
- Navigate this page:
- Similarities
- Differences
- REFERENCES
- ACKNOWLEDGEMENTS
Comparison to UMPRSJAM and UMPRS have a common ancestry and, not surprisingly, are similar in a number of ways. The publicly available version of UMPRS15 has not changed much in the last several years while, in contrast, we have been active in adding functionality to JAM. As a result, JAM is significantly more advanced than UMPRS in several areas. In this section, we describe the key similarities and differences between UMPRS and JAM. Similarities
Differences
Bratman88:Plans Michael Bratman, David Israel, and Martha Pollack. Plans and resource bounded practical reasoning. Computational Intelligence, volume 4, pages 349-355, 1988. Cohen87:Intention Philip R. Cohen, and Hector J. Levesque, "Intention = Choice + Commitment", Proceedings of the Sixth National Conference on Artificial Intelligence, Seattle, Washington, 410-415, July, 1987. Fikes71:STRIPS R. E. Fikes and N. J. Nilsson, "STRIPS: A New Approach to the Application of Theorem Proving to Problem Solving, Artificial Intelligence Journal, Vol 2, pgs 189-208, 1971. Georgeff87:Reactive Michael Georgeff and Amy L. Lansky. Reactive Reasoning and Planning. In Proceedings of the Sixth National Conference on Artificial Intelligence, pages 677-682, Seattle, Washington, August 1987. Georgeff86:Procedural Michael P. Georgeff and Amy L. Lansky. Procedural Knowledge. Proceedings of the IEEE Special Issue on Knowledge Representation, 74(10):1383-1398, October 1986. Gray97:Agent R. S. Gray, D. Kotz, G. Cybenko, and D. Rus. Agent Tcl. In W. Cockayne and M. Zyda editors, Mobile Agents, Manning Publishing, 1997. Huber93:UMPRS Marcus J. Huber, Jaeho Lee, Patrick Kenny, and Edmund H.Durfee, UM-PRS Programmer and User Guide, The University of Michigan, Ann Arbor MI 48109, 1993. [Available at http://members.home.net/marcush/IRS] Ingrand92:Architecture Francois Ingrand, Michael Georgeff, and Anand Rao. An Architecture for Real-Time Reasoning and System Control. IEEE Expert, 7(6):34-44, December 1992. Ingrand90:Managing Francois. F. Ingrand and Michael P. Georgeff. Managing Deliberation and Reasoning in Real-Time AI Systems. In Proceedings of the 1990 DARPA Workshop on Innovative Approaches to Planning, Scheduling, and Control, pages 284-291, San Diego, CA, November 1990. Lange98:Programming Lange, Danny and Mitsuru Oshima, Programming and Deploying Java Mobile Agents with Aglets, ISBN: 0-201-32582-9, 1998. Lee94:UMPRS Jaeho Lee, Marcus J. Huber, Edmund H. Durfee, and Patrick G. Kenny. UM-PRS: An Implementation of the Procedural Reasoning System for Multirobot Applications. In Conference on Intelligent Robotics in Field, Factory, Service, and Space (CIRFFSS’94), pages 842-849, Houston, Texas, March 1994. Lee94:Structured Jaeho Lee and Edmund H. Durfee. Structured Circuit Semantics for Reactive Plan Execution Systems. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 1232-1237, July, 1994. Lee96:Structured Lee, Jaeho, Structured Circuit Semantics, Ph.D. Thesis, Ann Arbor, MI, 1996. Lee97:Explicit Jaeho Lee. An Explicit Semantics for Coordinated Multiagent Plan Execution, Ph.D. dissertation, University of Michigan, 1997. Myers97:User User Guide for the Procedural Reasoning System, K. L. Myers, Technical Report, Artificial Intelligence Center, Technical Report, SRI International, Menlo Park, CA, 1997. Myers97a:Act The Act Editor User's Guide , K. L. Myers and D. E. Wilkins, Technical Report, Artificial Intelligence Center, SRI International, Menlo Park, CA, 1997 Myers97b:Act K. L. Myers and D. E. Wilkins, "The Act Formalism", Version 2.2, SRI International Artificial Intelligence Center, Menlo Park, CA, September 1997. Peine97:ARA Peine H. “ARA – Agents for Remote Action”, in William Cockayne and Michael Zyda editors, Mobile Agents, Manning Publishing, 1997. Wilkins95:Common D. E. Wilkins and K. L. Myers, "A common knowledge representation for plan generation and reactive execution," Journal of Logic and Computation, vol. 5, number 6, pp. 731--761, December 1995
A number of individuals have submitted bug reports, fixes, and extensions that have resulted in improvements to the JAM distribution. We would like to acknowledge these people here. Chip McVey from Johns Hopkins University’s Applied Physics Lab (JHU/APL) supplied several primitive functions for performing Lisp-like list manipulations (first, rest, last, etc.) on strings. Martin Klesen from DFKI (Germany) provided feedback on a number of areas that resulted in bug fixes (rest primitive action), and functional and efficiency improvements (recvMessage primitive, parsing, error handling rather than exiting to facilitate embedding JAM within application code). In addition, Martin submitted a useful primitive function (unassign). Robert Goldman, Kurt Krebsbach, X, and Y from Honeywell Inc. found a number of bugs, including the POST action (fixed) and mis-handling of variables in plan contexts when they are passed in as arguments in the goal specification (not fixed). We would also like to acknowledge the support of the institution of JHU/APL, which sponsored some extensions under a DARPA (Defense Advanced Research Projects Agency) subcontract. 1 The version number is a complex number representing a unit vector, reflecting that a certain amount of the “real” work has been done but that a certain amount still “imaginary”.☺ The version number will swing towards the “real” axis as the JAM intelligent agent architecture matures and more of the TODO items (please read through the “TODO” file in the distribution to see the long list of extensions we would like to eventually address) are implemented. The version number will never reach 1.0 + 0.0i as there will always be more to do. JAM releases may also be referenced by their release date, in this case, November of 2001. 2 Selection of one plan from all those that are applicable is performed either by metalevel plans or by a search for the highest utility plan in the set if metalevel plans are not applicable or do not exist. We discuss this in more detail later. 3 The jam directory also contains a jam.ico file that may be used in Windows-based PCs as an icon that can be associated with files with a .jam extension. 4 We are considering adding optional typing support. While untyped languages are great for prototyping systems, they are notoriously prone to runtime errors due to type mismatches that might have been caught using typing information at parsing/compile time. 5 It is meant only as an example of syntax and as an introduction to some of JAM’s plan components and actions is not meant to represent a truly functional plan. 6 For example, the same message-handling plan might be instantiated twice when a low-priority message is suspended because a high-priority message comes in and needs processing 7 See Section 10 for more details about the JAM interpreter execution cycle. 8 Given the current execution semantics, there is little need, if any, to ever use an and branch. 9 See Section 10 for more details about the JAM interpreter execution cycle. 10 Note that this differs from the retrieve action, which always ignores variable bindings and overwrites such bindings with value(s) from the world model. 11 This functionality will be implemented as a separate thread eventually. At this point in time, however, both the interpreter and the Observer are part of the same thread. 12 We include this argument for accessibility when writing metalevel primitives and similar functions which typical agent programmers will not usually write. 13 This last form, implementations of the PrimitiveFunction interface, can be found in the “primitives” directory that is a subdirectory of the primary source-code distribution 14 Metalevel plans may subgoal just as any other plan. There may therefore be “metalevel” plans that are invoked as subgoals and will therefore have goal specifications. 15 The public version is UMPRS V3.1. ORINCON Corporation has a proprietary version, nominally called V3.4, that has some extensions not found in V3.1. Directory: IRS IRS -> Regulating railway safety in an environment characterised by change IRS -> The state education department IRS -> The state education department IRS -> An architecture for Access to a compute Intensive Image Mosaic Service in the nvo IRS -> The state education department Download 391.89 Kb. Share with your friends:
|