Stanley J. Rosenschein. Plan synthesis: A logical perspective. In Proceedings of the Seventh International Joint ConferenceonArtificial Intelligence, pages 331--337, Vancouver, B.C., Canada, 1981. 195  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... e.g. reasoning about the concurrent execution of actions [ Lin and Shoham,1992; Baral and Gelfond,1993 ] sensing actions [ Levesque,1996 ] and exogenous events [ De Giacomo et al. 1997a ] In this paper we follow the approach to the representation of dynamic systems based on PDLs [ Rosenschein,1981 ] and its extension based on the correspondence with Description Logics [ De Giacomo et al. 1996; 1997b ] In fact, we extend such an approach in order to address action planning in the presence of sensing actions, concurrent actions and exogenous events, in the realm of Cognitive Robotics, ....

Stanley J. Rosenschein. Plan synthesis: A logical perspective. In Proc. of the 7th Int. Joint Conf. on Artificial Intelligence (IJCAI'81), pages 331-- 337, 1981.

....presented in this paper has never been proposed before. So far, failure has had very little attention in formal theories of actions. This is the case of Dynamic Logic [10] as we discussed in Section The intuition , and, as a consequence, of the approaches to planning based on Dynamic Logic, e.g. [12]. This is also the case of the work described in [16] which shares with our research most of the underlying motivations. The idea that actions in the real world may fail is, of course, present in most of the work on reasoning about actions and plans. But none of them provides a theory where ....

S. Rosenschein. Plan synthesis: A logical perspective. In Proc. of the 7th International Joint Conference on Artificial Intelligence, pages 331--337, Vancouver, British Columbia, 1981.

....into account failures. Up to now, the languages to describe embedded system activities arise from two different approaches. The first approach starts from a theoretical point of view, the so called theories of actions. They are languages focused on what an action is [ McCarthy and Hayes, 1969; Rosenschein, 1981; Allen, 1984; Rao and Georgeff, 1991; Gelfond and Lifschitz, 1993; Lifschitz, 1993; Lesperance et al. 1994 ] Internal models are usually assumed to be good enough to predict all (or most of) the external events. Similarly, information acquired from the environment is often assumed to be ....

S. Rosenschein. Plan synthesis: A logical perspective. In Proc. of the 7th International Joint Conference on Artificial Intelligence, pages 331--337, Vancouver, British Columbia, 1981.

....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 ....

Stanley J. Rosenschein. Plan synthesis: A logical perspective. In Proceedings of the seventh International Joint Conference on Artificial Intelligence (IJCAI'81), pages 331--37, 1981.

.... since in generated models R is universal, and hence R ff R for arbitrary action terms (see Passy Tinchev, 1991) The semantic behavior of R facilitates the formulation of domain constraints, that is, formulas which are assumed to be invariant over every performance of every action (see Rosenschein, 1981; Ginsberg Smith, 1988a; Stephan Biundo, 1993; in the knowledgebase literature they are also called integrity constraints, see Katsuno Mendelzon, 1991) A more detailed study of this simple yet powerful construct is given in section 5, when we turn to applications of (extended) ....

....pp. 170 172) 5.2. Planning We are now going to demonstrate that NPDL is an appropriate tool to formalize concepts essential to planning. Planning is a discipline that concentrates on problems adherent to the specification, verification, and synthesis of plans (see Manna Waldinger, 1980; Rosenschein 1981; Georgeff, 1987b; Stephan Biundo 1993) Plan synthesis concerns the composition ( synthesis ) of a plan, usually a compound action term, to achieve some specified goal or goals. The process that checks if the proposed plan meets its specification, is called verification. By specification one ....

[Article contains additional citation context not shown here]

Rosenschein, Stanley J., 1981, "Plan synthesis: a logical perspective," In Proceedings of the Seventh International Joint Conference of Artificial Intelligence, Vancouver, 331-337.

....information is quite natural and intuitive. 3 Logics have been previously used in work on planning. In fact, perhaps the earliest work on planning was Green s approach that used the situation calculus [25] Subsequent work on planning using logic has included Rosenschein s use of dynamic logic [46], and Biundo et al. s use of temporal logic [10, 12, 11, 53] However, all of this work has viewed planning as a theorem proving problem. In this approach the initial state, the action effects, and the goal, are all encoded as logical formulas. Then, following Green, plans are generated by ....

Stanley J. Rosenschien. Plan synthesis: A logical perspective. In Procceedings of the International Joint Conference on Artifical Intelligence (IJCAI), pages 115--119, 1981. 76

....the use of regression rewriting, for the purpose of verifying certain properties of Golog specified complex tests, and for generating complex tests as conditional plans. Our presentation draws upon (Lesperance 1994) and (Reiter 2000) Other related approaches to conditional planning include (Rosenschein 1981; Manna Waldinger 1987; Lobo 1998) Consider the Golog complex test given above to help discriminate the space of hypotheses generated when a car won t start. To verify that it is an individual discriminating test, it is necessary to ensure that for at least one of the hypotheses H , Kwhether(H; ....

.... designed for efficiency, but the results can serve as the foundation for building more efficient regression based complex test planning methods, much as similar results have served as the foundation for relatively more efficient regression based planning methods (McDermott 1991; Lesperance 1994; Rosenschein 1981). In future work we will evaluate the extension of current state of the art planning techniques based on SAT and Graphplan, to address the planning problems raised in this paper (Weld 1999) Summary In this paper we presented results towards a formal theory of testing for dynamical systems, ....

Rosenschein, S. 1981. Plan synthesis: A logical perspective. In Proc. Seventh International Joint Conference on Artificial Intelligence (IJCAI-81), 331--337.

....game script : ibid] each situation has its own script. Scripts can be used more than once. The use of scripts or schemata is pervasive in the field of planning and natural language understanding, both for plan recognition [Carberry, 1988; Ramshaw, 1991] and for plan synthesis [Allen, 1987; Rosenschein, 1981; Stefik, 1981b; Dean et al. 1988; Vere, 1983; Tate, 1977; Wilkins, 1984] among others) A good discussion of the sociological motivations for scripts is found in [Suchman, 1987] ffl Hierarchical decomposition. Researchers such as Sacerdoti [Sacerdoti, E.D. 1973; Sacerdoti, 1975] ABSTRIPS, ....

Stanley Rosenschein. Plan synthesis: a logical perspective. In Proceedings of IJCAI, pages 331--337, 1981.

....planners. In these frameworks, plans can only be represented as sequences of actions (sequential composition of actions) Deductive planning systems based on various situation logics (Green, 1969; Waldinger, 1977; Kowalski, 1979; Manna Waldinger, 1980; Manna Waldinger, 1986) and dynamic logics (Rosenschein, 1981; Kautz, 1982) fall into this category. Linear planning is believed to be exponentially less efficient than nonlinear planning (Chapman, 1987) and has rarely been adopted since NOAH (Sacerdoti, 1975) Furthermore, because of this linearity restriction, concurrent actions and interval actions ....

....the non temporal component of an assertion a proposition type. In planning, assertions of proposition types can be interpreted over either time points, time intervals or both. In situation calculus (McCarthy Hayes, 1969) and some other formal languages for planning (Manna Waldinger, 1986; Rosenschein, 1981), assertions of proposition types are interpreted over situations. A situation is a complete snapshot of the world at some instant in time, and, therefore, the assertions of the proposition types are implicitly interpreted at time points. In the planning representation language given by ....

Rosenschein, S. J. 1981. Plan Synthesis: A Logical Perspective. In: Proceedings of the 7th International Joint Conference on Artificial Intelligence.

....at any time and will return the best result computed until then. In this paper, we will concentrate on this type of reasoning. Many types of reasoning in AI such as diagnosis, classification, design and planning are characterised in terms of logical entailment [Reiter, 1987; Green, 1969; Rosenschein, 1981] A general approach for constructing approximate problem solving behaviour is to use an approximation of the logical entailment for characterising such a problem type and see what conclusions can be drawn using approximate reasoning for a particular problem. This approach was already taken in ....

S. Rosenschein. Plan synthesis: a logical perspective. In IJCAI'81, pages 331--337, 1981.

....this paper we consider conditional planning from a more abstract point of view. Instead of extending existing algorithms that produce non conditional plans, we view conditional planning as an automated reasoning task, like in the pioneering work on planning by Green (1969) in deductive planning (Rosenschein, 1981), and in recent work on planning by satisfiability algorithms (Kautz Selman, 1992, 1996) Instead of using a very general framework like the first order predicate logic or a dynamic logic, we choose a logic that is sufficiently expressive for representing conditional planning but also restricted ....

Rosenschein, S. J. (1981). Plan synthesis: A logical perspective. In Hayes, P. J. (Ed.), Proceedings of the 7th International Joint Conference on Artificial Intelligence, pp. 331--337 Los Altos, California. William Kaufmann.

....world. Logics of action and change have two uses in planning. The first is as a tool for specifying and proving properties of plans and planning algorithms. The second is as a representation of states and actions. An early planner using logics of action and change in the second way is found in [16]. The state, or states, of the world are described by arbitrary boolean formulas, and actions are modeled by a set of axioms that relate the value of a formula after an action to the value of some other formula before it. Regression and progression operators are defined, and a linear planner is ....

....space that contain only suboptimal solutions, if any at all. It has been used in practically every regression planner since STRIPS. In this section we will examine ways of adapting this technique to the complex goals allowed in K Gamma IA scenarios. The result will be nearly identical to that of [16]. Let A be an action with n branches of the form p i e i , i = 1: n. The preconditions p i are all mutually exclusive, and we can, without loss of generality, consider the action to be complete since if it is not it is easy to add a new branch with precondition p = W i=1: n :p i and empty ....

[Article contains additional citation context not shown here]

Stanley J. Rosenschein. Plan synthesis: A logical perspective. In IJCAI81, pages 331--37, 1981.

....AI are based on transition systems. For example, both STRIPS like formalisms [18, 3] and Action Languages [22] are formalisms to compactly represent transition systems (of different generality) Furthermore, formalizations based on logics such as the Situation Calculus [34, 39] or Dynamic Logics [40, 13], are also tightly related to transition systems. As goal specification we adopt automata on infinite words: the desirable traces correspond to the language accepted by the automaton. This way of specifying goals is very close to adopting linear time temporal logic (LTL) 16] as goal specification ....

S. J. Rosenschein. Plan synthesis: A logical perspective. In Proc. of IJCAI'81, 331--337, 1981.

....only the intuitions behind this theory. This is done by explaining actions behaviours in terms of state transitions. This can be the basis for a formal semantics of different formal languages, e.g. languages based on the situation calculus (McCarthy and Hayes 1969) dynamic and temporal logics (Rosenschein 1981; Harel 1984; Harel et al. 1982) action description languages (Gelfond and V.Lifschitz 1993) This allows us to keep the discussion general and independent of the particular formalism which might be chosen. The theory takes into account the following facts: 1) actions may fail, since they ....

S. Rosenschein. Plan synthesis: A logical perspective. In Proc. of the 7th International Joint Conference on Artificial Intelligence, pages 331-- 337, Vancouver, British Columbia, 1981.

....lavoro si propone di usare un sistema per la rappresentazione della conoscenza di tipo generale basato sul formalismo delle logiche descrittive. Ci o e possibile poich e, partendo dalla formalizzazione di un sistema dinamico in termini di una teoria di una logica dinamica proposizionale (PDL) [11], si pu o sfruttare la corrispondenza tra tali logiche e le logiche descrittive (DL) 14, 5] per ottenere una rappresentazione in DL del sistema dinamico. Nel caso specifico e stato usato CLASSIC [3] un sistema generale per la rappresentazione della conoscenza nel formalismo delle DL. Le logiche ....

....o che la costruzione di questo grafo richiede una conoscenza completa del sistema dinamico. Per modellare conoscenza incompleta sul sistema si pu o utilizzare una formalizzazione tramite assiomi, come ad esempio la PDL o il situation calculus [10] Seguendo l approccio descritto da Rosenschein in [11] si possono suddividere gli assiomi della teoria che descrivono il sistema in due sottoinsiemi: i) assiomi statici (o vincoli del dominio) che rappresentano la conoscenza sull ambiente e che non dipendono dallo stato in cui il sistema si trova, ii) assiomi dinamici, che descrivono le ....

S. Rosenschein. Plan synthesis: a logical perspective. In Proc. of the 8th Int. Joint Conf. on Artificial Intelligence, 1981.

.... of regular events which corresponds to the action modality of dynamic logic, whereas ordinary temporal logic cannot express this class of events at all [ Wolper, 1983 ] It is also interesting to note, in the AI literature, that Rosenschein s formulation of planning problems in dynamic logic [ Rosenschein, 1981 ] omitted the ff construct ) Despite its expressiveness, dynamic logic has been criticized within the AI community for its rigid change of state semantics [ Shoham and Goyal, 1988 ] More closely related to our approach is the family of process logics proposed by Pratt [ Pratt, 1979; Harel et ....

S. Rosenschein. Plan synthesis: A logical perspective. In Proceedings of the Seventh International Joint Conference on Artificial Intelligence, pages 331--337, 1981.

....plans in which the outcome of the agent s action may affect the next action taken by the agent. Theoretical work on this issue is mainly devoted to aspects of reasoning about knowledge and action (Moore, 1980; Halpern, 1988; Morgenstern, 1987) and to the logical formulation of conditional plans (Rosenschein, 1981). Specific mechanisms to construct conditional plans in which observable events and tests are explicitly declared are discussed as well (Wellman, 1990) These as well as the more classical work on conditional plans (Warren, 1976) and work that followed and extended it in various directions (Peot ....

Rosenschein, S. (1981). Plan Synthesis: A logical Perspective. In Proceedings of the 7th International Joint Conference on Artificial Intelligence.

.... foundations and guarantees for the resultant planners (highlighted by troublesome problems such as the Sussman anomaly and the register exchange problem) This led to a renewal of efforts in the 80 s to find viable approaches to wellfounded planning, exemplified by novel algorithms such as BIGRESS [22, 16] (based on dynamic logic) Bibel s linear connection method for plan generation [3] TWEAK [4] Chapman s partial order planner based on his modal truth criterion ) and SNLP [18] another systematic partial order planner using propositional STRIPS operators) These efforts have gained ....

S. Rosenschein. Plan synthesis: a logical perspective. In Proceedings of the Seventh International Joint Conference on Artificial Intelligence (IJCAI-81), pages 331--337, 1981.

....world to a few number of facts, or to multiply the number of operators. All these limitations of models based on the STRIPS assumption are well known, and have led to the definition of models of actions with context dependent effects, as in the Waldinger s planner [33] the Rosenschein s planner [31], and more recently in Wilkins SIPE [34] or in Pednault s ADL [28, 27] The alternative we propose in this paper consists in distinguishing between two types of facts in the description of the world, in order to define such a model of actions. Some particular facts are directly asserted by ....

....leads us to add to the state description some literals like (loc a r1) near a c) Since these literals represent some properties of a state and can be deduced from it, we call them deduced facts. As we said, this natural distinction between facts does not really appear in classical planners [12, 31, 5, 34]. To be complete, this logical description of the world state must include all the relations between different literals and we can define two main types of relations. First, there must be some relations to define the deduced facts. In the example from the mobile robot domain, we may define some ....

S. Rosenschein. Plan Synthesis: A logical Perspective. In Proceedings IJCAI, 1981.

....[DW91] for a general discussion on related subjects) Our work is mostly related to work on conditional planning. Theoretical work on this issue is mainly devoted to aspects of reasoning about knowledge and action (e.g. Moo80, Hal88, Mor87] and to the logical formulation of conditional plans [Ros81] Specific mechanisms to construct conditional plans, in which observable events and tests are explicitly declared are discussed as well [Wel90] These works as well as the more classical work on conditional plans (e.g. War76] and works that followed and extended it in various directions [PS92, ....

S.J. Rosenschein. Plan Synthesis: A logical Perspective. In Proc. 7th International Joint Conference on Artificial Intelligence, 1981.

No context found.

Stanley J. Rosenschein. Plan synthesis: A logical perspective. In Proceedings of the Seventh International Joint ConferenceonArtificial Intelligence, pages 331--337, Vancouver, B.C., Canada, 1981. 195

No context found.

Stanley J. Rosenschein. Plan synthesis: A logical perspective. In Proceedings of the Seventh International Joint Conference on Artificial Intelligence, pages 331--37. Morgan Kaufmann, 1981. Reprinted in [3].

No context found.

Rosenschein, S.: Plan synthesis: a logical perspective. In: Proc. of the 8th Int. Joint Conf. on Artificial Intelligence. (1981)

No context found.

Stanley J. Rosenschien. Plan synthesis: A logical perspective. In Procceedings of the International Joint Conference on Artifical Intelligence (IJCAI), pages 115--119, 1981. 76

No context found.

Rosenschein, S. J. Plan synthesis: a logical perspective. Proceedings of the Seventh National Conference on Artificial Intelligence. St. Paul, MN; 331337 (1988).

First 50 documents

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