D. McDermott. A temporal logic for reasoning about processes and plans. In James Allen, James Hendler, and Austin Tate, editors, Readings in Planning, pages 436 -- 463. Morgan Kaufmann Publishers Inc., 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....beliefs and desires, are an important component of the folk psychological concepts of intelligence and agency, especially as these concepts are used in AI. Specifically, intentions and reasoning about intentions are a crucial part of many important subareas of AI e.g. planning [Georgeff, 1987; McDermott, 1982] plan recognition [Pollack, 1986] natural language understanding [Grosz Sidner, 1990] and multiagent systems [Singh, 1991c] Perhaps the salient property of future directed intentions is that they involve commitment on the part of agents. This view has been gaining ground in the philosophical ....

McDermott, Drew; 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6(2):101--155.

....of time. Time is inherent in the very notion of monitoring. Without regard to it, a monitoring system misses out one (if not the most) important aspect of clinical information. While this fact, although very obvious, has long been ignored by work in the field of deductive (medical) systems [McDermott 82] other disciplines such as signal processing have made it their very basis right from the beginning. Despite a recent tendency to adopt time related issues in all branches concerned with clinical monitoring, there is still not one, agreed upon way of doing it. Rather, representation and ....

....reality, and intervalbased recordings are piecewise abstractions of continuous time reality. From this point of view, notions of facts and events or equivalents prominent in many artificial intelligence approaches to modelling time (see, for example, Shoham 87] including a critical treatise of [McDermott 82] and [Allen 84] are rather abstract even variables commonly associated with events are naturally observed as continuous time. Example If a stroke is defined as the sudden obstruction of a cerebral blood vessel, then the event stroke at would be observed as the step function t 0 ....

DV McDermott (1982) "A temporal logic for reasoning about processes and plans", Cognitive Science 6, 101--155

.... is to use states for cataloguing what is true at each individual point in time, for all such points [Pernault 90] Hence, according to this new terminology, a state is a chronicle of all that is true, was true, and will be true of the world, from the beginning of time through to the end of time [McDermott 82] Since states can be interpreted as chronicles, state transitions can be interpreted as the initiation of an action or course of action at a particular point in time. In some situation, however the effect of executing actions simultaneously is the same as doing them sequentially, regardless of ....

D. McDermott. "A Temporal Logic for Reasoning about Processes and Plans", Cognitive Science, Vol. 6, pp 101-155. 1982.

....TO VISUALIZE TIME, TENSE AND ASPECT COMMENT VISUALISER LE TEMPS ET L ASPECT Gtrard LIGOZAT Michael ZOCK LIMSI, Langage Cognition B.P. 133, 91403 Orsay Cedex France courtier 61ectronique: ligozat limsi.fr, ou, zock limsi.fr Nous allons dtcrire ci dessous une extension d un systme interactif de gtntration de ptu:ases. Apres une description de SWlM, acronyme de See What I Mean nous allons discurer la faqon dont on communique ....

McDermott D.. A Temporal Logic for Reasoning about Processes and Plans, Cognitive Science 6, 101-155, 1982.

....like Mary has long blonde hair then we do want # to hold at descendants of t as well. In the literature on foundations of planning, propositions of the latter kind have been called properties or facts, while propositions like Mary travelled all the way from London to Vienna are known as events [2, 14, 18]. Such ontological distinctions often form an integral part of the definition of a planning formalism. In Allen s work [2] for example, the definition of the predicates holds and occur, which are used to distinguish between di#erent types of propositions, is central to the entire system. holds ....

D. V. McDermott. A Temporal Logic for Reasoning about Processes and Plans. Cognitive Science, 6:101--155, 1982.

....Next State Output 37 Df: schedule the timing or sequencing specifications for a plan. A schedule can be represented as a sequence of activities that are indexed by time or triggered by events. A schedule may specify that activities or goal events should occur within certain time constraints [19]. Or a schedule may simply specify the order in which activities should occur regardless of the time. For example, the sequence of program steps in a robot assembly task typically does not depend on clock time, but on events that occur as the assembly task proceeds. If a plan is represented as a ....

McDermott, D. (1982) "A Temporal Logic for Reasoning about Processes and Plans," Cognitive Science. 6, pp. 101-155

....views (like in QPT[2] and are described in a formalism called VDL for View Design Language. The component s representation is a tree structure whose nodes are concepts, some of them describing the actions that control the component s evolution. Our execution model is based on execution steps [6] that end with the calculation of the next state of the view by F. We thus reuse the semantics of Sandewall s ego world [10] and Gurevich s evolving structures [3] You may find the complete description and operational semantics of the language in [8,9] We implemented this model in Java1.1: a ....

D.V. McDERMOTT. Temporal Logic for Reasoning about Processes and Plans. Cognitive Science, 6:101-155, 1982.

....we de ned work with the view at the present time. In order to answer to temporal requests, we need to know the component s state at each step of the execution. This is impossible, due to the limited amount of memory in computers. We propose to use chronicles [9] initially introduced by McDermott [27], to register the event occurrences in the view, like in Qpt [12] For each event that occurs in the system, be it internal or external, we build a model of the situation in which it occurs and In VDL, events correspond to the actions performed by the component. 28 register it (the ....

D.V. McDERMOTT. Temporal Logic for Reasoning about Processes and Plans. Cognitive Science, 6:101155, 1982.

....in the rule: i: Now(i) i 1: Now(i 1) The above indicates that from the belief (at time i) that the current time is in fact i, one concludes that it now is the later time i 1. That is, time does not stand still as one reasons. Note that temporal logics [ Allen and Ferguson, 1994, McDermott, 1982, Rescher and Urquhart, 1971 ] also have a notion of past, present and future, but these do not change as theorems are derived. These are speci cation logics external to the reasoner. This contrasts strongly with the agent based on board character of active logic. Technically, an active logic ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101-155, 1982.

....we next examine related work in temporal logic. 1.2.2 Temporal Logic Formal reasoning about time and action is not new. A great deal of research has been devoted to this field. Perhaps the two most influential temporal formalisms are those of Allen ( Allen, 1984 ] and McDermott ( McDermott, 1982 ] In [ Allen, 1984 ] a logic which permits reasoning about time is developed. In it time intervals are the principal objects in the domain. Three basic entities are associated with time: properties, events, and processes. HOLD(p; i) where p is a property type (e.g. red) and i is an interval ....

....i) denotes the fact that process p occurred over interval i. A process is said to occur over an interval i iff it occurs over some subinterval of i. Having set up a way to handle temporal information, Allen, 1984 ] then proceeds to handle actions, causation, intentions, and plans. McDermott [ McDermott, 1982 ] constructs his theory using fact types and event types. Unlike Allen, McDermott uses time points as primitive. T (t; p) denotes the fact that fact type p holds at time t. OCC(t 1 ; t 2 ; e) denotes the fact that event type e occurred over the interval t 1 ; t 2 . McDermott, 1982 ] then uses ....

[Article contains additional citation context not shown here]

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....the means to express the temporal aspects of data. Artificial Intelligence (AI) community has investigated many temporal logics which can deal successfully with various types of temporal information [1] 2] 3] 4] Intense research has resulted to the development of Temporal Logics [1] [5], 6] 7] 8] which have found application in many domains, such as, planning, temporal databases, verification of concurrent systems, VLSI design, etc. Many practical systems which implement Temporal Logics have also been reported. However, most of them are implementations of Modal temporal ....

McDermott D (1982) A Temporal Logic for Reasoning about Processes and Plans. Cognitive Science, 6, pp.101-155.

....entities, time points (i.e. instants) are often adopted. Intervals are then represented by their upper and lower temporal bounds (start and end time points) In practice, most systems employed in medical informatics applications have used a time point based approach, similar to McDermotts ###### [67], rather than use time intervals as the basic time primitives, as proposed by Allen [64] Several variations exist. Thus, Shahar [25] defines a set of time primitives, called ###########; predicates, however, such as values of clinical parameters, can only be interpreted over ##############, which ....

McDermott DV, A temporal logic for reasoning about processes and plans. Cognitive Science 1982; 6(2): 101155.

....the expressive power of the temporal logic. We show how incomplete relative temporal relations can be represented using TAND and AND connectives. Introduction Much work has been done to represent and reason temporal knowledge in a wide range of disciplines; including artificial intelligence[i][2][3] 4] 5] 6] databases[7] 8] 9] program verification[10] and software requirements specification languages[11] 12] 13] Temporal knowledge representation models introduced so far can be grouped into two: namely point based and interval based temporal models. In point based models time is ....

....employs HOLD(t,p) and IN(tl,t2) predicates in his temporal logic. HOLD(t,p) is true iff property p holds during time interval t while IN(tl,t2) is true iff time interval tl falls within time interval t2. Peter[5] uses and unary operators to represent start and end points of an event. McDermott[2] uses Lisp like statements while Manna and Pnueli[10] apply several first order temporal operators for program verification. However Torsun[16] says that first order temporal logic is not complete. Temporal operators available in first order temporal logic[10] 16] 17] do not provide enough ....

. McDermott, D. (1982) A Temporal Logic for Reasoning About Processes and Plans. Cognitive Science, 6, 101-155.

....diagnosis. Temporal reasoning has been considered in philosophy and logic since Thales and Zeno [19] however, it is only in the last two decades that temporal reasoning has been explicitly considered in artificial intelligence. McDermott and Allen, with their work in the early nineteen eighties [2 4,20], brought temporal reasoning into the AI mainstream. Other models for temporal reasoning include point algebras [32] semi intervals [10] temporal constraint networks [9] and weak representations of interval algebras [18] McDermott provides one of the earliest temporal representations [20] In ....

....[2 4,20] brought temporal reasoning into the AI mainstream. Other models for temporal reasoning include point algebras [32] semi intervals [10] temporal constraint networks [9] and weak representations of interval algebras [18] McDermott provides one of the earliest temporal representations [20]. In his approach, time is divided into a series of states with each state having an associated date, i.e. point in time. Facts are expressed as being true during particular states. Allen introduced interval temporal reasoning to the AI community [2, 4] Allen s interval algebra is governed by ....

McDermott, Drew V. "A Temporal Logic for Reasoning About Processes and Plans," Cognitive Science , 6 :101--155 (1982).

....to this as a partial path within the Situation Calculus tree. A path in the Situation Calculus tree corresponds to an infinite sequence of situations that start in S 0 , a partial path is a prefix of infinitely many paths. Thus, the notion of path corresponds to McDermott s notion of chronicle[19]. Any situation s identifies a unique partial path that starts at S 0 and ends with s. In fact, any situation in the domain S is fully identified with a sequence of actions. For instance, the sequence c 1 ; c 2 ; c n uniquely identifies the situation Sn , where: Sn = do( c 1 ; c n ....

....developments which obey all occurrence constraints imposed by the occurrence axioms. An essential aspect of our approach is that we rely on the notion of history. This notion is equivalent to the notion of chronicle in Drew McDermott s Temporal Logic for Reasoning about Processes and Plans [19]. To sum up, in this paper we: 1. Build upon a first order solution to the frame problem in the Situation Calculus. 2. Extend the language with explicit references to actions in histories. 3. Use histories to define external occurrences. 4. Define different notions of occurrences that can be ....

MCDERMOTT, D. A Temporal Logic for Reasoning About Processes and Plans. Cognitive Science 6 (1982), 101--155.

....cation are unique for ease of exposition. 2 The use of the credulous skeptical terminology to characterize these two broad reasoning strategies was rst introduced in Touretzky et al. 25] but the distinction is older than this; it was noted already by Reiter, and was described in McDermott [16] as the distinction between brave and cautious reasoning. 32 3 Reiter provides a proof procedure, sound and complete under certain conditions, for determining whether a formula is believable in this sense on the basis of a default theory. A di erent interpretation of this second credulous ....

D. McDermott, \A Temporal Logic for Reasoning about Processes and Plans" Cognitive Science, 6 (1982), 101-155.

....of logical omni science has received attention in the epistemic logic literature. It concerns the difficulty with the classical Hintikka possible world semantics [Hin62] that the agent always knows the logical consequences of his beliefs. Work in temporal logic involves reasoning about time (e.g. [All84, McD82, MS87]) Time is not treated as a crucial resource that must be carefully rationed by the agent, as it is spent in every step of reasoning. Moreover, most of the research till today assumes that the knowledge of the agent is consistent. 6 Step logics were introduced as a formal apparatus to model an ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....of logical omniscience has received attention in the epistemic logic literature. It concerns the difficulty with the classical Hintikka possible world semantics [Hin62] that the agent always knows the logical consequences of his beliefs. Work in temporal logic involves reasoning about time (e.g. [All84, McD82, MS87]) Time is not treated as a crucial resource that must be carefully rationed by the agent, as it is spent in every step of reasoning. Moreover, most of the formal models assume that the knowledge of the agent is consistent. Active logics (originally known as step logics ) were introduced as a ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....kind of resource limitation that is most evident in commonsense reasoners is the passage of time while the reasoner reasons. There is not necessarily any fixed and final set of consequences with which such a reasoning agent ends up. In a sense, this is a problem of modeling time. See [Allen, 1984, McDermott, 1982] Yet these treatments deal with reasoning about time, as opposed to reasoning in time. Reasoning in time refers to the fact that, as the reasoner reasons, time passes, and this passage of time itself must be recognized by the reasoner. Step logic is proposed as an alternative to the approaches ....

....process of deduction; there is no final state of reasoning. There are many examples of situations in which the effort or time spent making deductions is crucial. Consider Little Nell who has been tied to the railroad tracks. A train is quickly approaching. Dudley must save her. See [Haas, 1985, McDermott, 1982] It is not appropriate for Dudley to spend hours figuring out a plan to save Nell; she will no longer need saving by then. Thus if we are to model Dudley s reasoning, we must have a mechanism that takes into account the passage of time as the agent is reasoning. The Three wise men Problem is ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....occurrence. This kind of modeling that is called event calculus is advantageous because the frame problem [30] is avoided. Several researchers have made propositions to overcome the frame problem and one approach is to restrict the truth of single propositions to a temporal interval. In [2] [31], and [36] several models of this type were proposed. My model follows mainly those of Pelavin and Allen [36] in the temporal representation. The modal approach that is described later has some similarities with McDermott s logic. I distinguish between the prototype of an occurrence and the ....

McDermott, D., A Temporal Logic for Reasoning about Processes and Plans, Cognitive Science 6, pp. 101-155, 1982.

....DGvA99] Code, documentation, test results, and a full technical report are available on the web at http: www.cs.sfu.ca jim Time . 2 Related Work Our interests in this paper lie first with qualitative algebraic approaches to temporal reasoning. Other approaches include modal temporal logics [Ben83, McD82], in which modal operators may be used to express that a proposition is always or sometimes true in a past or future time. In an approach such as the situation calculus [MH69, LPR98] where reasoning about actions is formalised in first order logic, temporal precedence is implicit in the nesting ....

D.V. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....for cutting short action a after t seconds, if it would otherwise 34 have taken longer. The properties of truncated actions can be axiomatized using a Time function on situations which is real valued and one to one on any set of situations constituting a possible history of the universe (cf. McDermott 1982, Allen 1984) Trunc allows us to say, for example, that at all situations s 00 during W alk(R; x; y) starting in situation s and ending in s 0 , the uent formula moving toward(R; y; s 00 ) holds. Moreover, a slight generalization of explanation closure axioms allows us to extend ....

D. McDermott, \A temporal logic for reasoning about processes and plans," Cog. Science, 6:101-155, 1982.

....( LossOfSkill, PatientA ) 11 March) may be interpreted as follows: the fact (treated as a single entity ) that Patient A suffers form a loss of skills occurs at 11th of March. Some of the most influential approaches to temporal logics in AI, namely the works of Allen [1] McDermott [9], Kowalski and Sergot [8] can be subsumed under the category of reified temporal logics. Reification may be understand as a trick to achieve higher order expressiveness while staying at a first order representation. The cost to pay is a somehow unclear semantics, with respect to the reified ....

McDermott, D.: A temporal logic for reasoning about processes and plans. Cognitive Sci. 6 (1982) 101-155.

....mention only a few. For analyzing such systems, a number of time models have been proposed. Examples include temporal logics corresponding to branching time [7, 4] and partially ordered time [2, 10] The branching time model has been especially successful in its application to planning problems [6, 11]. Since most of the previous research on nonlinear time has concentrated on modelling properties 1 A problem is said to be tractable if it is solvable in polynomial time. 2 and theoretical foundations, computational aspects have received relatively little attention. There are a few papers that ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982. 26

....to facilitate knowledge engineering that we will describe later in this paper. A temporal ontology is based on a temporal logic. There are many different time theories in the literature [1] and there are difficulties associated with each of them. For example, instant based time theories [5,6,7] are not natural for representing events that take time, and Allens interval based theory has trouble with the Dividing Instant Problem [8] In [9, 10] the approach of treating both instants and intervals as independent primitives is presented, which is the approach we adopted. It is of interest ....

D. McDermott; A Temporal Logic for Reasoning about Processes and Plans; in Cognitive Science, 6:101-155, 1982

....Absolute theories take time elements, such as instants or intervals, as primitive, and regard events (occurrences, or processes, etc. as happening in time. These theories can be further divided into two types. One type treats instants of time as primitive, as exemplified in Newton (1934) and McDermott (1982). The other type takes intervals of time to be elementary, e.g. Newton Smith 1980, Hamblin (1972) Dowty (1979) and Allen and Hayes (1985, 1989) Relational theories of time, on the other hand, hold that our concept of time is derived (abstracted) from our experiences of events and thus regard ....

McDermott, D. 1982. A temporal logic for reasoning about processes and plans. Cognitive Science, vol. 6, pp. 101--155.

....to mention only a few. For analyzing such systems, a number of time models have been proposed. Examples include temporal logics corresponding to branching time [6, 4] and partially ordered time [1, 9] The branching time model has been especially successful in its application to planning problems [5, 10]. Since most of the previous This research has been supported by the ECSEL graduate student program. This research has been supported by the Swedish Research Council for the Engineering Sciences (TFR) under grant 97 301. research on nonlinear time have concentrated on modelling properties ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....mean that the referenced individual will acquire smallpox. The use of counterfactuals avoids problems with applying such a definition 9 This phenomenon exists whenever one modifies some event description with some temporal or spatial modifier. 15 simply in terms of a branching temporal logic [ McDermott, 1982 ] Cases involving events and not actions become problematic: if the future and past is fully determined by some set of causal laws then there is no branching temporal structure. Consequently, some counterfactual analysis is necessary (particularly in order to choose the closest world in which ....

....the door to the room. The following report (31) The guard kept the prisoner in the room by not leaving the area. is satisfactorily explained by definition 10.1: if the guard had left the area then the prisoner might have escaped. 13 This is also sometimes referred to as protecting a fact [McDermott, 1982]. 21 11 Rational Agency: Causal connections between mind and action Much of the impetus for belief, desire, and intention (BDI) models of rationality 14 has sprung from a certain brand of folk psychological wisdom: rendering behavior intelligible amounts to identifying the reasons for which ....

McDermott, Drew 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6:101--155.

....used to guide a planning search and reduce the potentially huge search space to a manageable size [18, 5, 13, 22] But the work on abstraction assumes that a plan is built without time constraints, so it is not applicable to problems that require incremental results. The work on temporal planning [1, 15, 17, 20] provides formalisms that represent the dynamic and continuous nature of a changing environment better than the discrete Strips style representation. But again these approaches fail to build plans in a useful way for time stressed situations. We have developed an approach to incremental planning ....

D. McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....This problem is very acute as soon as one takes seriously the fact that events answer to many descriptions. 2. 4 Other belief change operations and approaches to counterfactual reasoning An alternative means of defining counterfactuality is by way of a branching temporal logic [ von Wright, 1967; McDermott, 1982; Elkan, 1996 ] where a forward branching structure emerges as a consequence of an assumption of non determinism or free will on the part of agents being modeled: for example, some state, s, might represent the root of two branches, ff and fi, where each stands for the performance of some distinct ....

....that specify that a subtype of an event that occurred also occurred 7 . The formula 2W OE will mean that OE is true in all possible worlds (at the given time) Finally, the connective is the usual one for counterfactual 7 See the end of this section. The representation is essentially that of McDermott [1982] and also that of Cohen and Levesque [1990] in which propositional fluents are true at a point and events occur over intervals, with the difference that the duration or interval information is taken to be part of the event description in HL . 14 dependence: OE is meant to express the fact ....

McDermott, Drew 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6:101--155.

....Consider the statement, 2) The vaccine prevented him from getting smallpox. Unintended would be a suggestion that smallpox would have inevitably eventuated had the person not received the vaccine. A more acceptable definition in terms of a model of branching time was suggested by McDermott [ McDermott, 1982 ] e prevents e 0 just in case before e, e 0 was possible, while after e, e 0 became impossible. In this paper, I will refer to this definition of prevention as necessary prevention, for reasons that will become clear shortly. Sentence (2) is an example of necessary prevention. The problem ....

McDermott, Drew 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6:101--155.

.... among computer scientists because it allows the representation of a deterministic past (what happened did happen) together with an open future (all that could happen) The semantics regarding the topology of time is usually reflected by the temporal modalities provided for describing events [Mca1, BPM1, EHa1, Mcd1, MPn2]. Branching time raised a dispute on whether it should be interpreted as branching of time or branching in time, i.e. if branching is a structural property of time, or it is the course of events which branches in a linearly structured time. The second view has been generally accepted because of ....

....moved researchers in Artificial Intelligence to deal with time representation, it was the possibility of building systems which could reason about temporal facts which stimulated the most fruitful practical approaches. The application domains to which time reasoning has been applied are: planning [Mcd1, DMD1], Decision Support Systems [DPW1] Fault Diagnosis [MPe1] office systems [MPB1] deductive and temporal databases [KSe1, SAh1] The work of J. F. Allen must be acknowledged for its attempt to set the foundations for time reasoning systems [All1, All2, All3] he is concerned with properties of ....

[Article contains additional citation context not shown here]

McDermott D. -- A Temporal Logic for Reasoning about Processes and Plans -- Cognitive Science, vol. 6, pp. 101-155, 1982

....the world. We can think of the set of all possible histories of the world as forming a branching tree structure. A number of possible semantics and axiom sets based on such structures are discussed in (Thomason 1984) Branching structures have also been proposed as a basis for AI reasoning systems (McDermott 1982, Shoham 1988) but this approach does not seem to have been widely adopted. However, in the theory of computation, branching time logics, such as CTL (Emerson and Halpern 1986) have been extensively studied (Moller and Birtwistle 1996, Moller and Rabinovich 1999) 3 It can be argued that any ....

McDermott, D.: 1982, A temporal logic for reasoning about processes and plans, Cognitive Science 6, 101-155.

....support whatsoever nor any deontic support. Actions and time: have been extensively studied by many researchers in several areas of computing (e.g. Manna and Pnueli 1992; Haddawy 1991; Morgenstern 1988; Allen 1984; Lamport 1994; Nirkhe, Kraus, Perlis, and Miller 1997; Dean and McDermott 1987; McDermott 1982; Rabinovich 1998; Singh 1998) We present here the main di erences between others work and ours, and discuss work that combines time with deontic operators. Surveys of research on temporal reasoning include (Benthem 1991; Benthem 1995; Baker and Shoham 1995; Lee, Tannock, and Williams 1993; ....

McDermott, D. (1982, December). A temporal logic for reasoning about 56 processes and plans. Cognitive Science 6, 101-155.

....spilling the coffee by tipping the saucer or to cases of unintentionally spilling the coffee while involved in some other activity 1 . Further, 1) also makes This research was supported by the following grants: ARO no. DAAL 03 89 C 0031 Prime and DARPA no. N00014 90 J1863. 1 Though (McDermott 1982) presents an alternative representation of events by which one can, for example, represent actions such as running around the track, that work does not address the issues discussed here of generation reference to an attempt at performing a negative action (not spilling the coffee) where the ....

....attempting to maintain within some spatial interval. Periodically, someone pulls on either side: if I feel a tug to the left, I pull to the right, and vice versa. In each case, my action is of a distinct type: a left pull versus a right pull. These properties can be captured by the following: 7 (McDermott 1982) discusses an analogous notion of protecting a fact but leaves it as an open problem. Di Eugenio 1993) discusses maintenance actions in the context of instructions. done(i; t; m(OE) j (5) 9ff:done(i; t; ff) 3B later( OE) 9d8fi[happens(fi ; OE) f(fi; OE) d] done(i; t; ff) ....

McDermott, D. 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6:101-- 155.

....formulae, but only on atoms, making the formulae less clear and more awkward to reason with. A common approach to avoid this problem is to shift the atemporal objectlevel formulae into terms and introduce a special predicate Holds (also commonly called True) Examples of this approach appears in McDermott [1982] and Kowalski and Sergot [1986] The previous example could then look as follows: Holds(t 1 ; t 2 ; and(p(x) q(x; y) This is a so called reified logic (the object level formulae are reified, i.e. made into terms) The advantage is that we can qualify non atomic formulae temporally while ....

Drew McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

....look can be used either as a goal of achievement or a maintenance goal. The programming constructs required to do this coercion are discussed below. This typology of actions is found in work by [Vendler, 1967; Kaelbling, 1988; Dean and Wellman, 1991] having been developed in work by [McDermott, 1982; Allen, 1984; Israel et al. 1991; Moore, 1985] Note that in this approach, unlike that of Goldman, individual actions are goal oriented. This is consistent with a task oriented approach. The temporal aspect of these actions is quite different than that seen in the approach by Goldman et al. In ....

Drew V. McDermott. A Temporal Logic for Reasoning About Processes and Plans. Cognitive Science, 6:101--155, 1982.

....can be done, however, only if the maximal terminal moment for each action instance is defined i.e. it is definite. Uniqueness of initiation (C3) has the same effect for when the action instances begin. Together, these entail that the action instances do not overlap and have definite boundaries. McDermott [1982] noted that actions do not overlap, but was less clear about the existence of limits for the intensions of actions. Interestingly, constraint C4, which fills in the intensions of actions with all suffixes of periods that are in the intension, is satisfiable with constraints C2 and C3. However, ....

....definitions specialize to each case properly. This is useful since our models must often function at multiple levels of abstraction. We also discovered that several constraints must be stated on models to capture the intuitive notion of basic actions. The sole traditional constraint of no overlap [McDermott, 1982] says too little. Intuitively, acceptable models for the formal language must be branching to capture the branching time operators and to enable us to represent agents choices. Such models do not presuppose a tree structure in the technical presentation, because an equivalent presentation can be ....

McDermott, Drew; 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6(2):101--155.

....approximate a commitment with an agent s telling what it will do. The only way this can succeed is if we insist that agents do not come to this conclusion in anyway other than byhaving a commitment#intention. Otherwise, the agentmay fall victim to the #Little Nell problem #Cohen Levesque 1990a; McDermott 1982# in which the agent never forms the intention to act because it already believes the action will happen. But, if the conclusion that the agenteventually acts 3 We are aware that numerous researchers are attempting to add new performatives, many in the commissive class. Our point is that the ....

McDermott, D. 1982. A temporal logic for reasoning about processes and plans. Cognitive Science 6#2#:101#155.

....support whatsoever nor any deontic support. Actions and time: have been extensively studied by many researchers in several areas of computing (e.g. Manna and Pnueli 1992; Haddawy 1991; Morgenstern 1988; Allen 1984; Lamport 1994; Nirkhe, Kraus, Perlis, and Miller 1997; Dean and McDermott 1987; McDermott 1982; Rabinovich 1998; Singh 1998) We present here the main di erences between others work and ours, and discuss work that combines time with deontic operators. Surveys of research on temporal reasoning include (Benthem 1991; Benthem 1995; Baker and Shoham 1995; Lee, Tannock, and Williams 1993; ....

McDermott, D. (1982, December). A temporal logic for reasoning about processes and plans. Cognitive Science 6, 101-155.

....some (meaning sometime) is applied as a subscript to entire intersections; the first of the two statements asserts that WR(1) and R(2) intersect sometime. Representations for temporal relations have been extensively studied in the AI literature and elsewhere (see, for example, Allen, 1983) (McDermott, 1982), and (van Benthem, 1983) We use the usual relation symbol for total orders ( and what can be defined in terms of it and equality to represent temporal relations among events that, at the current level of analysis, are thought of as happening at instants. Events thought of as occurring over ....

McDermott, D.V., "A Temporal Logic for Reasoning about Processes and Plans," Cognitive Science 6 (1982), pp. 101-155.

....which is expressive enough for the domain of interest but no more so, as both characterization of the domain and reasoning about it are more cumbersome with the exotic varieties. In the case of dialogue about an essentially static domain, a point based representation such as the situation calculus (McDermott 1982) is adequate, as it can be assumed that the process of change to an agent s belief state resulting from one utterance will not continue past the next utterance. A slightly more expressive temporal logic involving a dense representation of time is sufficient for modelling the process of dialogue or ....

McDermott, D. V. (1982). A temporal logic for reasoning about processes and plans. Cognitive Science 6, 101--155.

....he is often supposed to take. The stronger position (i.e. the one that FOL must be used as the representation language) has been adopted by others, for example Moore [28] Genesereth and Nilsson [13] and Nilsson [32] and which was adopted by McDermott for a number of years (e.g. 24, 25] and [26]) Of the various arguments in favour of FOL, I consider those of Hayes to be the clearest, the best articulated and, possibly, the most easily available: the critique that follows is, if anything, intended as a compliment to him. With these preliminaries out of the way, a consideration of the ....

McDermott, D., A Temporal Logic for Reasoning about Processes and Plans, Cognitive Science, Vol. 6, pp. 101-155, 1982.

No context found.

D. McDermott. A temporal logic for reasoning about processes and plans. In James Allen, James Hendler, and Austin Tate, editors, Readings in Planning, pages 436 -- 463. Morgan Kaufmann Publishers Inc., 1990.

No context found.

McDermott, D.V.: A temporal Logic for Reasoning about Processes and Plans. Cognitive Science 6 (1982) 101--155

No context found.

McDermott D., A temporal logic for reasoning about processes and plans, Cognitive Science, 6:101-155, 1982.

No context found.

MCDERMOTT, D. A temporal logic for reasoning about processes and plans. Cognitive Science 6 (1982), 101--155.

No context found.

D. McDermott. "A temporal logic for reasoning about Processes and Plans". Cognitive Science 6, pp101-155, 1982.

No context found.

Drew McDermott. A temporal logic for reasoning about processes and plans. Cognitive Science, 6:101--155, 1982.

No context found.

McDermott D. (1982) A temporal logic for reasoning about processes and plans, Cognitive science, vol. 6, pp. 101-155.

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