Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach to Classical Planning. Springer-Verlag, Berlin, 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....which has only one step take the medicine with a default duration of 1 minute. 4. Specify an interstep constraint to ensure that the medicine taking occurs within one hour of finishing breakfast. As each pre constructed plan fragment or constraint is added, the PM performs step merging [26,30], that is, it checks to ensure the consistency of the daily plan being constructed and resolves any conflicts. To do this, it uses the same techniques for consistency checking that are used during plan execution; these techniques are described in the next subsection. The same GUI can be used ....

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. New York: Springer.

....the PM. This plan may then be changed in one of three ways: i) by the addition of new activities 3 ; ii) by the modi cation or deletion of (constraints on) activities already in the plan; iii) by the execution of one of the planned activities. In the rst two cases, PM performs plan merging [21, 8, 20, 19]: to ensure that the change does not introduce a con ict. In the third case, it propagates the constraints a ected by activity execution, as described in the example above. To adequately represent the client plans, it is essential to support a rich set of temporal constraints: for example, we may ....

Qiang Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, New York, 1997.

.... DISA, University of Trento, Via Inama, 38100 Trento, Italy Luca Spalazzi 2 Istituto di Informatica, University of Ancona, Via Brecce Bianche, 60131 Ancona, Italy 1 Introduction This is a review of the book Intelligent Planning: A Decomposition and Abstraction Based Approach by Qiang Yang [45], a book focused on classical planning. In recent years, the classical approach to planning has been challenged by many authors (e.g. 1,2,9,12,18,19] The question is whether its fundamental assumptions and complexity make it of any use in real world applications. The arguments against classical ....

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer Verlag, Berlin, Germany, 1997. 13

....reasoning is introduced. 1.1 Historical Roots of Case Based Planning The historical roots of case based planning rely on classical planning and case based reasoning. Classical planning involves the generation of plans by state or partial plan space search in order to satisfy a given goal (see [ Yang, 1997 ] for a good introduction) According to this definition, classical planning has been proved and experimented intractable (e.g. see [ Chapman, 1987; Korf, 1985; Yang, 1997 ] As a consequence, plenty of methods have been proposed to try to reduce the computational cost. One of the most used ....

.... planning involves the generation of plans by state or partial plan space search in order to satisfy a given goal (see [ Yang, 1997 ] for a good introduction) According to this definition, classical planning has been proved and experimented intractable (e.g. see [ Chapman, 1987; Korf, 1985; Yang, 1997 ] As a consequence, plenty of methods have been proposed to try to reduce the computational cost. One of the most used methods is hierarchical planning (e.g. Newell and Simon, 1972; Sacerdoti, 1974; Wilkins, 1988 ] see [ Yang, 1997 ] for a good introduction) The basic 4 idea of ....

[Article contains additional citation context not shown here]

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach.

....be formulated as a Constraint Satisfaction Problem or CSP, with temporal features. The process must consider temporal constraints, resource usage, and causal links (preconditions and effects) There has been a great deal of research done on similar problems by the Artificial Intelligence community [14, 17, 20]. A number of formalizations have been developed for variations with more or less expressivity. The two that most closely match our problem are the Disjunctive Temporal Problem (DTP) and the Conditional Disjunctive Temporal Problem (CDTP) For solving DTPs we have developed and implemented a new ....

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, 1997.

....more general plans for templates, and to optimize plans during their execution. More general plans can be achieved by plan decomposition which allows bene ts like concurrency in plan generation, more e ective usage of reusable plan parts, multi agent plan execution, balancing achievable goals, etc [41]. Plan decomposition may be achieved in addition to analytical methods for example using inductive learning techniques. As an RKB can be viewed as a nite automaton, where concepts represent states and action agents transitions between them, the planning agent can use di erent automata induction ....

....Through the plan generalization we implement what is called hierarchical planning, which is used as a method to reduce the computational cost of planning. The idea is to distinguish between goals and actions of di erent degrees of importance, and to solve the most important problems rst [41]. Its main advantage derives from the fact that by emphasizing certain activities while temporarily ignoring others, it is possible to obtain a much smaller search space in which to nd a plan. A generalized plan can be considered as a skeletal plan for a complete plan. Between every element of a ....

Q. Yang. Intelligent planning: a decomposition and abstraction based approach, Springer-Verlag, 1997.

....with similar, although not identical, aspects of planning and plan adaptation. Differences in planning frameworks, representations, and application domains have led to the differences between these approaches and our own. Plan merging in classical planning (Tsamardinos, Pollack, Horty 2000; Yang 1997) is concerned with finding a global plan to achieve multiple goals by combining separate plans for each goal. In our framework, the notion of goal achievement (which is all or none) is replaced by utility (which is continuous) In our framework, plans that fail to achieve one or more goals would ....

....reduced utility, whereas in the classical approach, such plans would have 0 utility and others would have either equivalent or cost dependent utility. Resolving precondition postcondition dependencies, as in the classical strips representation, does not by itself lead to plans of maximal utility. Yang (Yang 1997) addresses optimization based plan selection, using a heuristic approach to find the minimal cost plan by merging separate plans. Once again, this is designed to find a plan that achieves multiple goals with minimal cost, as opposed to maximizing overall plan utility while allowing for partial ....

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer.

....that the plans in S are mutually consistent, if such a set of constraints exists. The algorithm has three phases. In the first, it employs a new data structure, a conditional simple temporal network (CSTN) to identify conflicts between the plans. Next, it uses an approach developed by Yang (1997) to suggest a potential resolution of the identified conflicts. Finally, the CSTN is again used to check whether the proposed resolution observes all the temporal constraints. We have implemented our approach, and we present preliminary experimental evidence about domain factors that ....

....1989; Kambhampati Hendler 1992; Hanks Weld 1992; Veloso Carbonell 1994) Our focus in this paper is on the process by which the new plan can be merged into a set of existing plans, so that the resulting set is consistent. The most well developed prior work on plan merging is that of Yang (1997, Chap. 7) whose algorithms apply to classical plans, i.e. plans expressed using the STRIPS representation. 2 Yang s central idea is to em 1 In the section on Related Work , we discuss the relationship between this claim and apparently related results (Nebel Koehler 1995) 2 Yang s book ....

[Article contains additional citation context not shown here]

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. New York: Springer.

....reactive robots is that agents must be able to react (quickly) to the environment. The representation in this paper is not at odds with reactive agents but rather emphasises how to reason about current actions based on thinking about the future. Classic planning work in AI (Fikes Nilsson 1971, Yang 1997) has ignored the second point. The idea behind classical planning is to make a linear plan based on assuming what the world is like, and to patch this plan or replan if execution monitoring says that the plan has not worked. However, for virtually every interesting domain there are no actions ....

Yang, Q. (1997). Intelligent Planning: A Decomposition and Abstraction-Based Approach, Springer--Verlag, New York.

....parts are solved seperately by problem solvers. Another possibility is hierarchical abstraction, in which problems are organized into levels of hierarchy. The hierarchy can be used by the problem solvers to make the planning process more effective (for more literature about planing algorithms see. [18, 218, 227, 230]) 5.2.2 The Production system as a model of human problem solving The SOAR state, operator and result model was developed to explain human problem solving behavior [133] It is a hierarchical production system in which the conflict resolution strategy is treated as another problem to be solved. ....

....12.22 11.84 p 0.003 0.001 Table 6.2: Required plan length for all 42 I1 I6 and D1 combinations. h0 h1 h2 mean 13.79 2.9 1.9 sdev 43.46 3.66 2.52 p 0.055 0.04 Table 6.3: Required backtracking steps for all 42 I1 I6 and D1 combinations. The task ID1 D1 represents the Sussman anomaly [201, 227, 230, 137]. The problem is considered an anomaly because you cannot do either of the necessary first actions without undoing it at a later point. If you first put block A on block B after putting block C on the table, you will have to take it off again so that you can move block B onto block C. If you put ....

Q. Yang. Intelligent Planning A Decomposition and Abstraction Based Approach. Springer-Verlag, 1997. BIBLIOGRAPHY 289

....value associated with non observability, but unpredictable exogenous events and uncertain action effects cannot be modeled consistently if these assumptions are adopted. For an overview of early classical planning research and the variety of approaches adopted, see [1] as well as the recent text [140]. Task oriented versus Process oriented Problems It is useful at this point to pause and contrast the types of problems considered in the classical planning literature with those typically studied within the MDP framework. Although problems in the AI planning literature have emphasized a ....

Qiang Yang. Intelligent Planning : A Decomposition and Abstraction Based Approach. Springer Verlag, 1998.

....This possibility of successive detailing of the plans is of great importance in order to support the formation of nested coalitions for the achievement of a goal, such as proposed in the approach now presented. Therefore, the proposed plan representation is based on a hierarchical structure [10][14], where the plans are characterized by collections of activities that, in turn, can be decomposed. Having this in mind it is possible to define the plans in a hierarchical way, which correspond to strategies of activity realization in order to reach a main goal. An activity can, in turn, ....

Yang, Q. - Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, 1997.

....plans. In the present paper, we focus on the reasoning that occurs after a set of conditions for ensuring compatibility has been computed. That is, we develop algorithms by which an agent can compute a cost estimate for the plan P in the context C. An algorithm for this purpose was developed in [29], using a dynamic programming approach to find an optimally merged plan, one with minimal cost. This cost is precisely (P [ C) which could then be combined with an exact value for (C) to yield an exact value for (P=C) The option P could then be accepted or rejected depending on the relation ....

....plan, one with minimal cost. This cost is precisely (P [ C) which could then be combined with an exact value for (C) to yield an exact value for (P=C) The option P could then be accepted or rejected depending on the relation between (P=C) and fi(P) We do not rely on the algorithm developed in [29], however, for two reasons. First, it applies only to classical plans, which lack metric temporal constraints. And second, it is not in general necessary to compute the exact value of (P=C) in order to evaluate the new option; instead, as suggested earlier, an agent may be able to accept or reject ....

Qiang Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, New York, 1997. 34

....ensure that the plans in S are mutually consistent, if such a set of constraints exists. The algorithm has three phases. In the #rst, it employs a new data structure, a conditional simple temporal network #CSTN#, to identify con#icts between the plans. Next, it uses an approach developed byYang #1997# to suggest a potential resolution of the identi#ed con #icts. Finally, the CSTN is again used to check whether the proposed resolution observes all the temporal constraints. Wehave implemented our approach, and we present preliminary experimental evidence about domain factors that ....

....1989; Kambhampati Hendler 1992; Hanks Weld 1992; Veloso Carbonell 1994#. Our focus in this paper is on the process by which the new plan can be merged into a set of existing plans, so that the resulting set is consistent. The most well developed prior work on plan merging is that of Yang #1997, Chap. 7#, whose algorithms apply to classical plans, i.e. plans expressed using the STRIPS representation. 2 Yang s central idea is to em 1 In the section on #Related Work , we discuss the relationship between this claim and apparently related results #Nebel Koehler 1995#. 2 Yang s book ....

[Article contains additional citation context not shown here]

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. New York: Springer.

....a condition. The OP description is limited to its abstraction level, regardless of lower level information; for example, the OP above does not clarify the valves positions in the context or effects fields. In this sense, the OP representation is incomplete, and corresponds to the relaxed models [ Yang, 1997 ] used in hierarchical planning. Even if a sequence of OPs seems correct for some level of abstraction, interaction problems between effects and expected preconditions may arise at lower ones. Although it may be possible for each OP to compile off line its extended effects and context fields, ....

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach, pages 163--171. Springer-Verlag Publisher, 1997.

....the mall. In traditional plan generation, this would be handled by a step re use function. When a new plan is added to an existing context of commitments, explicit consideration of step merging becomes crucial. For further discussion of step merging, including algorithms for efficient merging, see [Yang, 1997]. We associate the cost of an incomplete or unscheduled plan with the cost of the least expensive way in which it might be carried out; i.e. the cost of an arbitrary plan P is the cost of the least expensive fully scheduled, complete plan Q such that Q is a refinement of P , where refinements ....

Yang, Q. (1997). Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, New York.

....(quickly) to the environment. The representation in this paper is not at odds with reactive agents (Poole, Mackworth 2 Goebel 1998, Chapter 12) but rather emphasises how to reason about current actions based on thinking about the future. Classic planning work in AI (Fikes Nilsson 1971, Yang 1997) has ignored the second point. The idea is to make a linear plan based on assuming what the world is like, and to patch this plan or replan if execution monitoring says that the plan has not worked. However, for virtually every interesting domain there are no actions whose consequences can be ....

Yang, Q. (1997). Intelligent Planning: A Decomposition and AbstractionBased Approach, Springer--Verlag, New York.

....the process sketched in Section 3 of evaluating a new option P against the background of a context C might actually be accomplished. As explained earlier, the first step of the process is determining whether P is, in fact, strongly compatible with the context C. This problem has been studied in [22], which develops a constraint satisfaction procedure to determine plan compatibility. This algorithm, Combine, takes as input two plans which, in our case, would be P and C identifies all the threats between them, constructs a constraint satisfaction problem (CSP) representing the threat ....

....a generalization of it that can handle such plans [17] Our focus in this paper is on the second step in the reasoning process, in which the agent computes the cost estimate for plan P in context C. Given an option P that is compatible with C, we might next appeal to the Optimal Merge algorithm [22], which uses a dynamic programming approach to find an optimally merged plan, that is, one with minimum cost. This cost is precisely (P [C) which could then be combined with an exact value for (C) to yield an exact value for (P=C) The option P could then be accepted or rejected depending on ....

Qiang Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach. Springer, New York, 1997.

....stages. In the first, it employs a Conditional Simple Temporal Network (CSTN) to identify conflicts among plans. CSTNs extend Simple Temporal Networks (Meiri 1992) to include branching nodes. After conflicts have been identified, the second stage of our algorithm uses an approach developed by Yang (1997) to suggest a potential resolution of the identified conflicts. Yang s approach is not guaranteed to generate resolutions that are consistent with all the temporal constraints allowed in our constraint language, so the third stage of the algorithm involves re using the CSTN to check for ....

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. New York: Springer.

....algorithms that enable an artificial agent to perform the reasoning processes described above. The first stage of the processing involves determining whether a new option P is strongly compatible with a given context C. Algorithms for checking plan compatibility have been studied in earlier work (Yang 1997); we have extended the algorithms developed there to handle plans with quantitative temporal constraints. The second stage of processing computes the estimates ffl Gamma P[C and ffl P[C , comparing these to fi(P) and then iterating to refine the estimates if needed. The computation of the ....

Yang, Q. 1997. Intelligent Planning: A Decomposition and Abstraction Based Approach. New York: Springer.

....optional step. In planning, the pattern of flying to a major hub then to another major hub before reaching the final destination, is analogous to achieving plan subgoals. A major hub would correspond to a common subgoal in a plan space. Subgoal identification is an important problem in planning [8, 14]. Once mined, this knowledge could be used in some planners to reduce the search space and improve plan efficiency. In our approach, the common high level sequences can be mined using a sequential pattern mining technique, which may involve operations like merge , op tion (collapsing) etc. ....

Q. Yang. Intelligent Planning -- A Decomposition and Abstraction Rased Approach. Springer-Verlag, 1997.

....In the area of plan recognition, the objective is to uncover the intention and goal of a series of observations in actions. Examples are works by Allen [2] and Woods, et al. 13] There is a large body of work in plan generation on how to generate plans from action descriptions and goals [1, 14, 12]. An example is the search control rule learning methods [4, 10] Zaki, et al. 15] developed a technique for plan mining, which first performs a brute force expansion of all the nodes and then mines sequence patterns using the Apriori technique [11] on the data which indicate high incidence of ....

....an optional step. In planning, the pattern of flying to a major hub then to another major hub before reaching the final destination, is analogous to achieving plan subgoals. A major hub would correspond to a common subgoal in a plan space. Subgoal identification is an important problem in planning [8, 14]. Once mined, this knowledge could be used in some planners to reduce the search space and improve plan efficiency. In our approach, the common high level sequences can be mined using a sequential pattern mining technique, which may involve operations like merge , option (collapsing) etc. ....

Q. Yang. Intelligent Planning --- A Decomposition and Abstraction Based Approach. Springer-Verlag, 1997.

No context found.

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach to Classical Planning. Springer-Verlag, Berlin, 1997.

No context found.

Yang, Q.: Intelligent Planning: A Decomposition and Abstraction Based Approach, Springer, Berlin, 1997.

No context found.

Q. Yang. Intelligent Planning: A Decomposition and Abstraction Based Approach to Classical Planning. Springer-Verlag, Berlin, 1997.

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