Bacchus, F., & Kabanza, F. (1995). Using temporal logic to control search in a forward chaining planner. In Proceedings of the 3rd European Workshop on Planning.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....can be learned using genetic programming. The results indicate that policies can be learnt which not only solve all test problems, but which can also outperform policies coded by hand. The results generated here are comparable with systems which use hand coded control knowledge such as TLPLan [3], TALPlanner [9] and SHOP [12] We are now working to apply L2Plan to more planning domains and to refine the learning method used. In doing the former, we will modify the system to support PDDL with typing: this will restrict the space of possible rules and should make the learning task easier. ....

Bacchus, F. and Kabanza, F. (1996). Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New directions in AI planning, 141--

....and single agent domains should be carried out to investigate the complexity of NADL s representation of concurrent actions. Several planners, in particular prodigy [41] have shown that domain knowledge should be used by a planning system in order to scale up to real world problems. Also [1] show how the search tree of a forward chaining planner can be efficiently pruned by stating the goal as formula in temporal logic on the sequence of actions leading to the goal. In this way the goal can include knowledge about the domain (e.g. that towers in the blocks world must be built from ....

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New directions in AI planning, pages 141--153. ISO Press, 1996.

....have been implemented. The potential of these techniques is demonstrated by applying TALplanner to a number of standard planning benchmarks in the literature. Keywords: Planning, Temporal Logics, Action and Change, Knowledge Representation 1. Introduction Recently, Bacchus and Kabanza et al. [9 11,22] have been investigating the use of modal temporal logics to express domain specific search control knowledge for forward chaining planners. This approach, implemented in the TLplan system, has demonstrated impressive improvements in efficiency when compared to many recent state of the art ....

....plan: 12 J. Kvarnstrom and P. Doherty TALplanner #occ [0,1] pick(ball1, left) #occ [1,2] pick(ball2, right) #occ [2,3] move to(roomB) #occ [3,4] drop(ball1, left) #occ [4,5] drop(ball2, right) #occ [5,6] move to(roomA) #occ [6,7] pick(ball3, left) #occ [7,8] move to(roomB) #occ [8,9] drop(ball3, left) 3.2. Searching for Plans Although Figure 2 provides an abstract view of a planner for the TAL formalism, it provides no information as to how a suitable set of action occurrences should be generated, much less how it should be generated efficiently. The answer to this question ....

[Article contains additional citation context not shown here]

F. Bacchus and F. Kabanza, `Using temporal logic to control search in a forward chaining planner', in New Directions in AI Planning, eds., M. Ghallab and A. Milani, 141--153, IOS Press, (1996). Available at ftp://newlogos.uwaterloo.ca/pub/bacchus/BKEWSP96.ps.gz.

....powerful but should be extended to enable modelling of constructive synergetic effects as described in Section 4. Also, more experiments comparing multi agent and single agent domains should be carried out to investigate the complexity of NADL s representation of concurrent actions. As argued by Bacchus and Kabanza (1996), domain knowledge must be used by a planning system in order to scale up to real world problems. They show how the search tree of a forward chaining planner can be efficiently 49 pruned by stating the goal as formula in temporal logic on the sequence of actions leading to the goal. In this way ....

Bacchus, F., & Kabanza, F. (1996). Using temporal logic to control search in a forward chaining planner. In Ghallab, M., & Milani, A. (Eds.), New directions in AI planning, pp. 141--153. ISO Press.

....c #1999, American Association for Artificial Intelligence (www.aaai.org) All rights reserved. control knowledge in a purely declarative manner by encoding the control as additional constraints. A recent example of the e#ectiveness of declarative control knowledge is the TLPlan system by Bacchus and Kabanza (1996; 1998) In the TLPlan system, control knowledge is represented by formulas in temporal logic. For example, the next operator from temporal logic allows one to specify what can and cannot happen at the next time step. The control knowledge is used to steer a forward chaining planner. One of the ....

....planning methods. We hope that our findings can be used to further enhance these methods, and in general deepen our understanding of the design space of planning systems (Kambhampati 1997) Temporal Logic for Control In TLPlan the control knowledge is encoded in temporal logic formulas (Bacchus and Kabanza 1996; 1998) The best way to introduce this approach is by considering an example control formula: 2 (# [p : airplane(p) # [l : at(p, l) # [o : in wrong city(o, l) in(o, p) # # in(o, p) In temporal logic, 2f means f is true in all states from the current state on and #f means f is true in ....

Bacchus, F. and Kabanza, F. (1996). Using temporal logic to control search in a forward-chaining planner. In New Dir. in Planning, M. Ghallab and A. Milani (Eds.), IOS Press.

.... in the style of (Cesta Oddi 1996) as well as intermediate tasks, like in (Bacchus Kabanza 1996) or in the style of the Hierarchical Task Networks approach (Erol, Hendler, Nau 1994; Yang 1990) Furthermore, LTL can express domain knowledge useful to guide the search, as shown in (Bacchus Kabanza 1995). This latter issue is the motivation of this work. We believe that the possibility of specifying extra problem speci c information, such as the domain expert is often aware of, is of great importance. In fact, such additional knowledge can be of great help both in reducing the search space and ....

Bacchus, F., and Kabanza, F. 1995. Using temporal logic to control search in a forward chaining planner. In Proc. of the TIME-95 International Workshop on Temporal Representation and Reasoning.

....Previous research on these concepts has been shackled by poor models of time and action, thereby leading to unnecessarily restricted or spurious results. The logic CTL was designed over a decade ago for reasoning about programs. It has increasingly found its way into the AI literature, e.g. [Bacchus Kabanza, 1995; Rao Georgeff, 1991; Singh Asher, 1990; Wooldridge, 1995] Therefore, it is a good choice for a formal language for a conceptual framework. CTL has also been used in the specification of BDI systems. Rao Georgeff, 1991] develop a framework in which various properties of intentions can be ....

Bacchus, Fahiem and Kabanza, Froduald; 1995. Using temporal logic to control search in a forward chaining planner. In Proceedings of the International Workshop on Temporal Representation and Reasoning.

....bounds as O(n 2 ) for US, O(n 3 ) for both GN1 and GN2, and O(n 4 ) for MHS. The original [4] did not mention any bound better than O(n 3 ) for near optimal BW planning, though quadratic time implementations have been described by other authors such as Chenoweth [2] and Bacchus and Kabanza [1]. 2.2.1 How to make US linear The key to making US a linear time algorithm is to find a way to compute which blocks are in position in time proportional to the number of blocks, and to execute this computation only once in the course of the problem solution. We do this by means of a combination ....

F. Bacchus & F. Kabanza, Using Temporal Logic to Control Search in a Forward Chaining Planner, Proceedings of EWSP-95 (1995).

....Nonetheless, we see at least two reasons for examining it in more detail. In the first place, for good or ill, bw is by far the most extensively used example in the planning literature. It often serves for demonstrating the merit of domain independent techniques, paradigms and planners. See (Bacchus Kabanza 1995; Kautz Selman 1992; 1996; Schoppers 1994) for recent examples. In order to assess the benefits of these approaches and the significance of the claims formulated in the literature, it is therefore necessary to know certain basic facts about bw, such as what makes optimal 1 bw planning hard ....

....in time linear in the number n of blocks. This improves on the known complexity of these algorithms. The original (Gupta Nau 1992) did not mention any bound better than O(n 3 ) for near optimal bw planning, though O(n 2 ) implementations have been described by other authors (Chenoweth 1991; Bacchus Kabanza 1995). function INPOS (b : block) boolean if b = table then return true if not Examinedb then Examinedb true if Sib 6= Sgb then InPosition b false else InPosition b INPOS(Sib) return InPosition b procedure US ( INIT( for each b 2 Bnftableg do if Clearb then UNSTACK(b) for each ....

Bacchus, F., and Kabanza, F. 1995. Using temporal logic to control search in a forward chaining planner. In Proc. EWSP-95, 157--169.

....to be considerably faster and requires less memory. The TAL versions also permit the representation of durative actions with internal state. In proceedings: 6th Int l Workshop on Temporal Representation and Reasoning (TIME 99) IEEE, 1999. 1. Introduction Recently, Bacchus and Kabanza et al. [3, 4, 5, 11] have been investigating the use of temporal logics to express search control knowledge for planning. In the approach, they utilize domain specific control knowledge represented as formulas in a first order modal tense logic to effectively control a forward chaining planner. The approach is ....

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New Directions in AI Planning, pages 141--153. ISO Press, 1996.

....1 The main advantage of the use of LTL as a planning language derives from its rich expressive power and the underlying simple model of time and actions. LTL can easily been used to express domain restrictions in the style of [8] as well as domain knowledge useful to guide the search, as shown by [1]. In this work we show how to represent action centered planning problems in the style of STRIPS like languages, with the addition of intermediate tasks to be accomplished. Planning problems can be represented in different forms, each encoding making the model construction procedure simulate a ....

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In Proc. of the TIME-95 International Workshop on Temporal Representation and Reasoning, Melbourne, Florida, April 1995.

....some circumstances, alleviate the computational difficulties of dynamic programming. Search based approaches to solving MDPs can use heuristics, learning, sampling, and pruning to improve their efficiency [3, 6, 7, 8, 9] Declarative search control knowledge, used successfully in classical planning[2], might also be used to prune the search space. In an MDP, this could be viewed as restricting the set of policies considered. This type of approach has been explored in the more general context of value iteration for MDPs in, e.g. 11, 18] local policies or finitestate machines are used to ....

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab, A. Milani, eds., New Directions in Planning, pp.141--153, 1996.IOS Press.

....our planner is able to solve (Brafman Hoos 1999) its complexities to be useful. The main feature of the approach we have presented is the expressive language it provides to accomplish such modeling. Furthermore, we have found that the mechanism of expressing search control, first promoted in (Bacchus Kabanza 1996), remains feasible in planning domains that include resources. As our empirical results show, our system is capable of generating feasible solutions very quickly with such control knowledge. We have not yet been able to properly evaluate the quality of the solutions being generated (although they ....

Bacchus, F., and Kabanza, F. 1996. Using temporal logic to control search in a forward chaining planner. In Ghallab, M., and Milani, A., eds., New Directions in AI Planning. ISO Press, Amsterdam. 141--153.

No context found.

Bacchus, F., & Kabanza, F. (1995). Using temporal logic to control search in a forward chaining planner. In Proceedings of the 3rd European Workshop on Planning.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In Proc of the 3rd European Workshop on Planning (EWSP), Assisi, Italy, Aug. 1995. AAAI Press.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New Directions in Planning, pages 141-- 153. IOS Press, Amsterdam, 1996.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. New Directions in Planning, M. Ghallab and A. Milani (Eds.), pages 141-153, 1996.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. New Directions in Planning, M. Ghallab and A. Milani (Eds.), pages 141-153, 1996.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New Directions in Planning, pages 141--153. IOS Press, Amsterdam, 1996.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab, A. Milani, eds., New Directions in Planning, pp.141--153, 1996. IOS Press.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New directions in AI planning, pages 141--153. ISO Press, 1996.

No context found.

F. Bacchus and F. Kabanza. Using temporal logic to control search in a forward chaining planner. In M. Ghallab and A. Milani, editors, New Directions in Planning, pages 141-153. IOS Press, Amsterdam, 1996.

No context found.

Bacchus and Kabanza. Using Temporal Logic to Control Search in a Forward Chaining Planner. New Directions in Planning, M.Ghallab and A.Milani(Eds.)10S Press pages 141-153,1996.

No context found.

Bacchus F., and F. Kabanza, Using temporal logic to control search in a forward chaining planner, in New Directions in AI Planning, M. Ghallab and A. Milano (eds.), 141-153, IOS Press, 1996.

No context found.

Bacchus F. and Kabanza F. (1994), Using Temporal Logic to Control Search in a Forward Chaining Planner, unpublished.

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