Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint ConferenceonArtificial Intelligence, pages 888--900, Cambridge, MA, 1977.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....has been proven to be free from interference in S. The case WM is assumed to be globally available throughout the executor, and describes the state of the agent s environment during execution. 5 Related Work The current paradigm for plan generation begins with NOAH (Sacerdoti, 1975) and NonLin (Tate, 1977); this approach has been extended by others (Vere, 1981; Wilkins, 1984, Currie Tate, 1985) but the core idea remains the same. A planner searches through a space of incomplete plans, each of which is partially ordered. Each plan is incomplete in the sense that it must be further refined to be ....

Austin Tate. Generating Project Networks. In Proceedings of the Fifth Joint Conference on Artificial Intelligence, pages 888-893, Boston, MA, 1977. International Joint Committee on Artificial Intelligence.

....performs state based forward and backward planning with a goal agenda. The planning state consists of assumptions and (sub)goals. The planner searches for a solution, i.e. for a sequence of instantiated operators whose application infers the goal from the initial state. Similar to HTN planning [15], planning operators can be expanded. The recursive expansion of a plan yields a calculus level natural deduction proof [5] that can be executed which means can be checked for logical correctness. That is, the plan execution corresponds to a soundness check which is performed at the ....

....Skolem functions in the e ect part of the (preconditions, e ects) pairs. 3.4 Employing External Systems PDDL does not provide a connection to external systems. Some AI planning systems make use of experts [20] RAX PS [6] uses experts in the development of plan fragments. Moreover, Nonlin [15] employed so called compute conditions which were also used in an extended way as the interface to rich external systems in O Plan [12] In mega, the application conditions and the outline computations can establish interfaces to external expert systems. Various expert systems exist for ....

A. Tate, `Generating project networks', in Proc. of IJCAI-77, pp. 888-893. Morgan Kaufmann, (1977).

....skeletal and opportunistic [3] Non hierarchical planners use goals directly to find operators as in NOAH [4] and STRIPS [5] planners. In hierarchical planners, planning begins at an abstract level, but later abstract goals are expanded into more detailed subgoals as in ABSTRIPS [6] NONLIN [7] and DEVISER [8] Skeletal planners store successful earlier plans in a plan database. Before planning begins, goals are compared against the skeletal plans. If one or more plans in the database satisfy the current goals and the current world model, the best plan will be chosen as in MOLGEN [9] ....

A. Tate, "Generating Project Networks", Proceedings IJCAI-77, Cambridge, Massachusetts, 1977, pp. 888-893.

....by NSF Grants IRI 9306580 and EEC 9402384, and ONR grant N0001491 J 1451. Any opinions, ndings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily re ect the views of the National Science Foundation or ONR. noah(Sacerdoti, 1975) nonlin (Tate, 1977), deviser (Vere, 1983) and sipe (Wilkins, 1988) Until now, there has been very little analytical work on the properties of HTN planners. One of the primary obstacles impeding such work has been the lack of a clear theoretical framework explaining what a HTN planning system is, although two ....

....nding operators that has the desired e ects, and by asserting the preconditions of those operators as subgoals. On the other hand, one of the motivations for htn planning was to close the gap between AI planning techniques and operations research techniques for project management and scheduling (Tate, 1977). htn planners search for plans that accomplish task networks, and they plan via task decomposition and con ict resolution, which we shall explain shortly. A task network is a collection of tasks that need to be carried out, together with constraints on the order in which the tasks are carried ....

[Article contains additional citation context not shown here]

Tate, A. Generating Project Networks In Proceedings of IJCAI, 1977. pp 888|889.

....of search is determined by the number of possible modifications of the current plan. An incomplete plan may be represented in many different ways. It may be a totM order sequence of operators (as in Strips [Fikes and Nilsson, 1971] or a partial order plan (as in Tweak [Chapman, 1987] Non Lin [Tate, 1977], and SNLP [MeAllester and Rosenblitt, 1991] the operators of the plan may be instantiated (e.g. in NoLimit) or contain variables with codesignations (e.g. in Tweak) the relations between operators and the goals they establish may be marked by casual links (e.g. in Non Lin and SNLP) In ....

Austin Tate. Generating project networks. In Proceedings of IJCAI, pages 888- 893, 1977.

....architecture [Carbonell et al. 1990, Veloso, 1989] develops a method to solve problems nonlinearly that explores different alternatives at the operator and at the goal ordering levels. Commitments are made during the search process, in contrast to a least commitment strategy [Sacerdoti, 1975, Tate, 1977, Wilkins, 1989] where decisions are deferred until all possible interactions are recognized. With the casualcommitment approach [Minton et al. 1989] background knowledge, whether hand coded expertise, learned control rules, or heuristic evaluation functions, guides the efficient exploration ....

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint Conference on Artificial Intelligence, pages 888--900, 1977.

....plan) Problems occur when partial plans interact with one another. For example, plan segments satisfying subgoals A and B might work fine when taken in isolation, but they may conflict when taken together, preventing both goals from being simultaneously satisfied (see, e.g. 22] 27] [40], 41] Hierarchical planners are one solution to the difficulties encountered in these early systems (see, e.g. 26] 28] 37] Very briefly, planning is carried out at several levels of abstraction. High level or abstract plans are used to generate skeletal or outline solutions. As ....

A. Tate, "Generating Project Networks," Proc. Fifth Int'l Joint Conf. Artificial Intelligence, pp. 888-893, 1977.

.... as the process of preserving the list of negated clipped or declipped literals (causal links) as part of the abductive process (as we will se bellow) Most of the systems from the AI planning literature represents plan assessment knowledge in one way or another way (with the exception of NONLIN [24]) However, there is no harm on having different representations for the same knowledge in a planner. Task method decomposition in the AEC planner Using the planning library we will characterize the problem solving method employed by the AEC planner. Although the specification of a theory of ....

A. Tate. Generating Project Networks. In Proceedings of IJCAI-77, pages 888-893, Boston, 1977.

....achieve a goal, and then adding further actions by detailing the action specified by that net. Unfortunately, because NOAH could not backtrack, it would in some situations commit actions that could lead it into dead ends. The next task network based planner was the NONLIN planner of Austin Tate [57], which was based loosely on NOAH. It extended the idea of critics from NOAH into the concept of an expert . Experts provided guidance to the nodes in the task network as to how to detail their actions if choices were available. Since backtracking was permitted in order to avoid the ....

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint Conference on Artificial Intelligence, pages 888--893, 1977.

....[Fikes and Nilsson, 1971] Robinson, 1965] In addition to being lifted, most modern planners are nonlinear they maintain a partial order on plan steps rather than a total order. This partial order is gradually refined during the planning process [Sacerdoti, 1975] Sacerdoti, 1977] [Tare, 1977], Chapman, 1987] Finally, some planners use abstraction spaces in which planning is first done at a high level of abstraction and then low level details are filled in once a high level plan has been found [Sacerdoti, 1974] Korf, 1987] Yang and Tenenberg, 1990] Nonlinear planners are ....

Austin Tate. Generating project networks. In IJCAI77, pages 888- 893, 1977.

....[Fikes and Nilsson, 1971] Robinson, 1965] In addition to being lifted, most modern planners are nonlinear they maintain a partial order on plan steps rather than a total order. This partial order is gradually refined during the planning process [Sacerdoti, 1975] Sacerdoti, 1977] [Tate, 1977], Chapman, 1987] Finally, some planners use abstraction spaces in which planning is first done at a high level of abstraction and then low level details are filled in once a high level plan has been found [Sacerdoti, 1974] Korf, 1987] Yang and Tenenberg, 1990] Nonlinear planners are ....

Austin Tate. Generating project networks. In IJCAI77, pages 888- 893, 1977.

....is not easy at least for large scale problems with fast job throughput. The idea, then, is to have the planner create a rough plan first, and as the abstract tasks get closer to execution, refine it stepwise toward a fully specified plan. Related work includes [4] in the scheduling area and [7, 8, 9, 11] in the planning area (although the latter approaches do not consider the temporal refinement aspect) This technique is very useful because it gives the planner the ability to take a high level view and potentially saves handling low level details for distant tasks for which the requirements are ....

Tate, A. 1977. Generating Project Networks. In Proceedings of the Fifth International Joint Conference on Artificial Intelligence (IJCAI-77), 888--893.

....An alternative to the data driven approach is goal driven search, in which a single agent searches for a way to satisfy its goals in the game. Goal driven search has been extensively explored in the Arti#cial Intelligence literature, in particular as hierarchical task network (HTN) planning [42, 15]. When multiple agents need to be modelled and can compete against one another, this approach becomes adversarial planning.In this paper we describe how adversarial hierarchical task network planning can provide a framework for goal directed game playing in Go. We consider di#erent types of goal ....

....This section describes an adversarial planning architecture which models goal driven reasoning for adversarial domains. The goal driven approach and use of abstract plans is motivated by work on hierarchical task network (HTN) planning. HTN systems were #rst used in NOAH [34] and INTERPLAN [42] and have since been extensively studied in the AI planning #eld. Erol et al. 15] give a complete de#nition for an HTN scheme and present UCMP, which is a provably sound and complete HTN planner and provides a good template for this type of system. 2.1. Principles of HTN planning HTN planning ....

A. Tate, Generating project networks, Proc. 5th Internat. Joint Conf. on Arti#cial Intelligence (IJCAI'77), 1977.

....to totally ordered) sequences of actions was first introduced by Sacerdoti [Sac75] and used in NOAH. This created a shift in how the planning is viewed: from state space search to searching the space of (possibly incomplete) partially ordered plans. However, NOAH is incomplete. Tate s NONLIN [Tat77] fixed some of the problems, but the full formal account of partial order plans has not been available until [Ped86] In his landmark paper [Cha87] Chapman clarified many issues not addressed by previously existing planning research and presented a provably sound and complete partial order planning ....

R. Tate. Generating Project Networks. In Proceedings of the Fifth International Joint Conference on Artificial Intelligence (IJCAI-77), pages 888--893, 1977.

.... Formal analyses of partial order and hierarchical planning algorithms define some desirable properties of plans, such as justifiability and correctness [Kambhampati, Knoblock, Yang, 1995; Yang, 1990; Tate, 1996] Generative planners such as SIPE, NOAH,andNONLIN [Wilkins, 1988; Sacerdoti, 1977; Tate, 1977] use critics that detect problems in the plans that they generate while planning. Much of the work on planning does not exploit background knowledge and ontologies, which we believe is crucial technology to advance the state of the art in process modeling. We have developed a tool that checks ....

Tate, A. 1977. Generating project networks. In International Joint Conference on Artificial Intelligence.

....world is created to check that these externally satis ed preconditions will indeed come about. If any are endangered or any undesirable events are detected they are resolved by modifying the a ected parts of the current plan. In the current work the planner is represented by Tate s Nonlin ([17]) which can be regarded as having two major parts: a set of schemata which form a World Model, albeit a very limited one; and a Plan Generator. Nonlin s world model is limited in that it cannot model continuous events, make use of events occurring external to the plan, model resources and their ....

....of di erent strategies from simply re executing a failed action to dynamically creating its own operator schemata, tailored to the speci c failure context from which repair plans can be generated. The 22 operators of the excalibur are de ned using the Task Formalism (tf) description language ([17]) The new schemata are created using a technique called Dynamic Schema Generation (dsg) which allows the planner to take the operator schemata de ned by the user and to edit them to create new schemata capable of solving the problem with which it is faced. The plan reasoner can only make check ....

A. Tate. Generating project networks. In Proceedings of the Fifth International Joint Conference on Articial Intelligence, pages 888{ 893, Los Altos California, 1977. William Kaufmann Inc.

....sequence is a goal state. Planning is sometimes refered to as domain independent planning, since the domain, as described by the available operators and their effects, is a part of the problem instance. Even though the problem has been much studied since the beginnings of artificial intelligence [13, 11, 6, 27, 29, 30, 19, 26, 22, 23, 21, 20, 24, 2], the most common formalisation is still the same as was used in the STRIPS system [11] The STRIPS formalism requires operators to have conjunctive preconditions and deterministic context independent effects. A later development, ADL [23] extends STRIPS to incorporate context dependent effects ....

Austin Tate. Generating project networks. In Proceedings of the fifth International Joint Conference on Artificial Intelligence (IJCAI'77), pages 88--93, 1977.

....a reference to the pitch, nominal diameter and the owner screw will be stored in this feature. 4. 3 Process Planning and Plan Evaluation Module To perform process planning for microwave module designs, weuse an approach from artificial intelligence called hierarchical task network (HTN) planning [11, 28, 35, 39]. We have also used this approach in some of our other work [33] HTN planning proceeds by taking a complex task to be performed and considering various methods for accomplishing the task. Each method provides a way to decompose the task into a set of smaller tasks. By applying other methods to ....

Tate, A. 1977. Generating project networks. In Proc. 5th International Joint Conf. Artificial Intelligence, 888--893. Morgan Kaufmann, San Mateo, CA.

....the total order restriction, to produce a sound and complete forwardsearch approach to HTN planning that has the same expressive power as unrestricted HTN planning. Overview of HTN Planning The basic ideas used in HTN planning were originally developed more than 20 years ago (Sacerdoti 1977; Tate 1977). To create plans, HTN planning uses task decomposition, in which the planning system decomposes tasks into smaller and smaller subtasks until primitive tasks are found that can be performed directly. HTN planning systems have knowledge bases containing methods. As shown in Figure 1(a) each ....

Tate, A. 1977. Generating project networks. IJCAI-77, 888--893.

....plan generation. The key challenge is efficiently retrieving the most relevant plan and then adapting it to fit the current context. Our focus is on plan repair via merging, and the retrieval process (as well as the merging process) is based on expected utility. Plan repair in classical planning (Tate 1977; Wilkins 1988) is concerned with finding replacements for plan wedges whose conditions have failed. Although our goal structure resembles plan wedges, we are concerned with finding a repair that maximizes the global utility, rather than concentrating on restoring conditions of the remainder of ....

Tate, A. 1977. Generating project networks. In Proceedings of IJCAI-77, 888--893. IJCAI.

....Sussman s HACKER system [58] However, the critics in NOAH still made choices among alternative orderings choices which may later turn out to have been wrong. This made NOAH incapable of solving a number of problems, since it didn t save the choice points for backtracking. Tate s NONLIN planner [59] reintroduced backtracking search to the planning problem. Over the next few years, additional enhancements were made to the basic nonlinear planning algorithm. For example, Vere s DEVISER [63] considered planning in time, i.e. some actions must occur within a time window, and Wilkins SIPE [66] ....

Tate, A., 1977. "Generating project networks", International Joint Conference on Artificial Intelligence (IJCAI-77), pp. 888-893.

....the plan repairer and credit assignment. 5. The newly generated recipe is stored into plan library. Other Case Based Planners Kambhampati s work on the PRIAR reuse framework [Kam89, KH89, KH92] applies a case based approach in a classical planning framework. PRIAR is an extension of NONLIN [Tat77a, Tat77b] which allows the planning systems to annotate plans being created with information about the dependency structure between operators in the completed plan. This information can also be used to guide retrieval, reuse, and re planning. R. Oehlmann, D. Sleeman, and P. Edwards [OSE92, OSE93] suggest ....

Austin Tate. Generating Project Networks. In Proceedings of the Fifth Annual International Joint Conference on Artificial Intelligence, Cambridge, Massachusetts, U.S.A., August 22 -- 25 1977. 98

....An alternative to the data driven approach is goal driven search, in which a single agent searches for a way to satisfy its goals in the game. Goal driven search has been extensively explored in the Artificial Intelligence literature, in particular as Hierarchical Task Network (HTN) planning [42,15]. When multiple agents need to be modelled and can compete against one another, this approach becomes adversarial planning. In this paper we describe how adversarial hierarchical task network planning can provide a framework for goal directed game playing in Go. We consider different types of goal ....

....applies data driven search for local tactical lookahead. 6 More details (including a precise breakdown of the planning algorithm) can be found in [45] 6 of abstract plans is motivated by work on hierarchical task network (HTN) planning. HTN systems were first used in Noah [34] and Interplan [42] and have since been extensively studied in the AI planning field. Erol et al. 15] give a complete definition for an HTN scheme and present Ucmp, which is a provably sound and complete HTN planner and provides a good template for this type of system. 2.1 Principles of HTN Planning HTN planning ....

A. Tate. Generating Project Networks. In Proceedings of the 5th International Joint Conference on Artificial Intelligence (IJCAI'77). 1977.

No context found.

Tate, A. (1977) Generating Project Networks, Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI-77), pp. 888-893, Cambridge, MA, USA, Morgan Kaufmann.

No context found.

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint ConferenceonArtificial Intelligence, pages 888--900, Cambridge, MA, 1977.

No context found.

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint ConferenceonArtificial Intelligence, pages 888--900, Cambridge, MA, 1977.

No context found.

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint ConferenceonArtificial Intelligence, pages 888--900, Cambridge, MA, 1977.

No context found.

Tate, A., 1977. Generating project networks, IJCAI-77, 888-893.

No context found.

Austin Tate. Generating project networks. In Proceedings of the Fifth International Joint Conference on Artificial Intelligence, pages 888--893, 1977.

No context found.

Austin Tate. Generating project networks. In Proceedings of the 5th IJCAI, pages 888-- 893, 1977.

No context found.

Austin Tate. Generating project networks. In Proc. International Joint Conference on Artificial Intelligence (IJCAI), 888-893, 1977.

No context found.

Austin Tate. Generating project networks. IJCAI-5, page 888 to 893, 1977.

No context found.

Austin Tate. Generating project networks. In IJCAI77 [32], pages 888-- 93. Reprinted in [3].

No context found.

Austin Tate. Generating project networks. IJCAI-5, page 888 to 893, 1977.

No context found.

A. Tate. Generating project networks. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI), pages 888--893, 1977. 12

No context found.

Austin Tate, Generating project networks, Proceedings of the International Joint Conference on Arti cial Intelligence (IJCAI) , 1977, 888-893.

No context found.

Austin Tate, Generating project networks, Proceedings of the International Joint Conference on Arti cial Intelligence (IJCAI), 1977, 888893.

No context found.

A. Tate. Generating project networks. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI-77), pages 889--900, 1977.

No context found.

A. Tate. Generating project networks. In IJCAI87, Proceedings of the Tenth Int. Joint Conference on Artificial Intelligence, 1987.

No context found.

A. Tate. Generating Project Networks. In Proceedings of IJCAI-77, pages 888-- 893, Boston, MA, 1977.

No context found.

A. Tate. Generating Project Networks. 1977.

No context found.

A. Tate. Generating Project Networks. In Proceedings of IJCAI-77, pages 888--893, Boston, MA, 1977.

No context found.

Austin Tate. Generating project networks. In Procceedings of the International Joint Conference on Artifical Intelligence (IJCAI), pages 888--893, 1977.

No context found.

A. Tate. Generating project networks. In Proceedings of IJCAI-5, pages 888--893, Boston, MA, 1977.

No context found.

A. Tate. Generating project networks. In Proceedings IJCAI-77, pages 888-- 889, 1977. Cambridge, Massachusetts.

No context found.

A. Tate. Generating project networks. In Proceedings of 5th InternationalJoint Conference on Artificial Intelligence, pages 888--893, 1977.

No context found.

A. Tate. Generating Project Networks. In Proceedings of IJCAI-77, pages 888--893, Boston, MA, 1977.

No context found.

A. Tate. Generating project networks. Proceedings of the Fifth International Joint Conference on Arti cial Intelligence, pages 888-893, 1977.

No context found.

Tate, A., "Generating Project Networks", Proceedings of the International Joint Conference on Articial Intelligence, pp. 888-89, 1997.

No context found.

Tate, A. 1977. Generating project networks. IJCAI-77.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC