Knight K.: Are Many Reactive Agents Better Than a Few Deliberative Ones? In: Proc. IJCAI '93, Chambery, 1993, 132--137  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the problem solver, rather than from the original initial state. These algorithms are different than ours since they assume that a node can be expanded in the computer s memory even without an actual agent physically visiting this node. Also, these algorithms are designed for large graphs. Knight [9] has presented a multi agent version of RTA called MARTA . In MARTA every agent runs RTA [10] independently. Kitamura et al. 8] have modified MARTA by using coordination strategies based on attraction and repulsion for MARTA . These strategies are employed only in tie breaking situations. ....

K. Knight. Are many reactive agents better than a few deliberative ones. In Proc. of IJCAI-93, pages 432--437, 1993.

.... techniques that are related to cooperative search have been proposed for solving hard computational problems, including: sequential restart strategies with diverse heuristics [52, 41, 31, 10] and parallel independent search with stochastic search algorithms [45, 40, 34, 36, 29] Knight [35] compares the performance of a system of many co 4 PARKES HUBERMAN operative agents with simple search heuristics to a system of a few agents with more complex search heuristics, and Aldous and Vazirani [1] describe a cooperative search technique called Go with the winners . Game theorists ....

Kevin Knight. Are many reactive agents better than a few deliberative ones? In Proc. 13th Int. Joint Conf. on Artificial Intelligence (IJCAI-93), pages 432--437. AAAI Press, CA, 1993.

....time of execution. In our case, the task is to construct the (parallel) algorithm providingthe best possible solution given a required utility. Simple parallelization, with no information exchange between the processes, may speedup the process due to high diversity in solution times. For example, Knight (1993) showed that using many reactive instances of RTA search (Korf 1990) is more beneficial than using a single deliberative RTA instance. Yokoo Kitamura (1996) used several search agents in parallel, with agent rearrangement after pregiven periods of time. Janakiram, Agrawal, Mehrotra (1988) ....

Knight, K. 1993. Are many reactive agents better than a few deliberative ones? In Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence, 432--437.

....reaches the goal by repeating the cycle of planning and execution. Real time search cannot guarantee to find an optimal solution, but can interleave planning and execution. Various extensions of real time search have been studied in recent years (Russell and Wefald 1991; Pemberton and Korf 1992; Knight 1993; Chimura and Tokoro 1994; Hamidzadeh and Shekhar 1995; Kitamura et al. 1996; Yokoo and Kitamura 1996) This paper focuses on extending real time search algorithms for autonomous agents and for a multiagent world. Though real time search provides an attractive framework for resource bounded ....

K. Knight, "Are Many Reactive Agents Better than a Few Deliberative Ones?," IJCAI-93, pp.432-437, 1993.

.... et al. 1993; Johnson et al. 1989; Boese et al. 1994) and parallel independent search that aims to take advantage of high variances in computation time for stochastic search algorithms (Rao and Kumar, 1993; Luby and Ertel, 1993; Kauffman and Levin, 1987; Kornfeld, 1981; Huberman et al. 1997) Knight (1993) compares the performance of a system of many cooperative agents with simple search heuristics to a system of a few agents with more complex search heuristics, and Aldous and Vazirani (1994) describe a cooperative search technique that they term Go with the winners . Schaerf et al. 1995) study ....

Knight, K. (1993). "Are many reactive agents better than a few deliberative ones?" in Proc. 13th Int. Joint Conf. on Artificial Intelligence (IJCAI-93), pp. 432--437.

....tasks in deterministic domains by moving a fictitious agent in the state space. Since fictitious agents move in almost no time, local search spaces that correspond to lookaheads of one or two action executions often minimize the sum of planning and plan execution time in this case (Korf 1990; Knight 1993). Agent centered search methods provide alternatives to traditional search methods, such as A (Nilsson 1971; Pearl 1985) They have, for example, successfully been applied to both traditional search (Korf 1990) and STRIPStype planning (Bonet et al. 1997) often outperforming traditional search ....

....an individual agent centered search each but sharing the search information,thereby reducing the planexecution time. For example, off line planning tasks can be solved on several processors in parallel by running a real time heuristic search method on each processor and let them share their values (Knight 1993). Similarly, exploration tasks can be solved with several robots by running an agent centered search method such as Method 2 on each robot and let them share the maps. While these properties can make agent centered search methods the planning methods of choice, it is important to realize that they ....

Knight, K. 1993. Are many reactive agents better than a few deliberative ones? In Proceedings of the International Joint Conference on Artificial Intelligence. 432--437.

....the power of real time search algorithms, which often outperform more traditional search techniques. Empirical results for real time search algorithms have typically been reported for domains such as ffl sliding tile puzzles (such as the 8 puzzle) Korf 1987; 1988; 1990; Russell Wefald 1991; Knight 1993; Korf 1993; Ishida 1995) and ffl gridworlds (Korf 1990; Ishida Korf 1991; Ishida 1992; Pemberton Korf 1992; Pirzadeh Snyder 1990; Thrun 1992;Matsubara Ishida 1994;Stentz 1995;Ishida 1995) Such test domains permit comparisons between search algorithms. It is therefore important that the ....

....search algorithm (min LRTA ) and compare its complexity in general state spaces to the most efficient and less efficient real time search algorithms. LRTA Type Search Korf s Learning Real Time A (LRTA ) algorithm (Korf 1990) is probably the most popular real time search algorithm (Ishida 1995; Knight 1993; Koenig Simmons 1995) The version we use here is closely related to Q learning, a widely used reinforcement learning method, see (Koenig Simmons 1996a) We call it LRTA with minimalistic lookahead (min LRTA ) because the search horizon of its action selection step is even smaller than that ....

Knight, K. 1993. Are many reactive agents better than a few deliberative ones? In Proceedings of the IJCAI, 432--437.

.... cost structures, including ones with non uniform costs (in a way analogous to [ Koenig and Simmons, 1992 ] where we assume deterministic state spaces) Furthermore, various methods have been proposed that improve the performance of LRTA , see for example [ Matsubara and Ishida, 1994; Ishida, 1993; Knight, 1993; Hamidzadeh, 1992; Ishida, 1992; Russell and Wefald, 1991; Shekhar and Dutta, 1989 ] we have applied and analyzed these methods in the context of min max LRTA . If one could make assumptions about nature s strategy (for example, if one knew that nature is a neutral coin flipper) or the state ....

Kevin Knight. Are many reactive agents better than a few deliberative ones? In Proceedings of the IJCAI, pages 432--437, 1993.

.... Levesque, Mitchell 1992; Luby, Sinclair, Zuckerman 1993; Johnson et al. 1989; Boese, Kahng, Muddu 1994) parallel independent methods (Rao Kumar 1992; Luby Ertel 1993; Kauffman Levin 1987; Kornfeld 1981; Huberman, Lukose, Hogg 1997) and cooperative parallel multiagent search (Knight 1993; Aldous Vazirani 1994; Hogg Williams 1993; Clearwater, Huberman, Hogg 1991) A general theory predicts superlinear speedup in the performance of individual agents when the search methods are diverse and the agents are able to utilize information found in other parts of the search space ....

Knight, K. 1993. Are many reactive agents better than a few deliberative ones? In Proc. 13th International Joint Conference on Artificial Intelligence (IJCAI-93), 432--437.

....[Eph94] Nau90] and others, each agent plans for a specific subtask, without regard to which agent or agents will execute it. Finally, there is disjunctive distribution, in which each agent works on the entire plan, searching a different part of the search space. This is somewhat like the work in [Kni93]. In the disjunctive approach, we may divide the search space by a complete but disjoint set of assumptions. If the assumptions are disjoint, then since only one of the assumptions can be true, only one of the plans being developed in parallel will be executed. Thus there is no need to resolve ....

Kevin Knight. "Are many reactive agents better than a few deliberative ones?" Proceedings of the 13th International Joint Conference on Artificial Intelligence (IJCAI), pp. 432-437, 1993.

....action executions is linear in the number of actions available in the current state. If this number does not depend on the size of the state space, then neither does the search time between action executions. Such methods have recently been proven to be very powerful if executed by multiple agents [12]. There is a potential trade off between exploitation and exploration when selecting an action, see [33] Exploitation means to behave optimally according to the current knowledge, whereas exploration means to acquire new knowledge. Exploration consumes time, but may subsequently allow the agent ....

Kevin Knight. Are many reactive agents better than a few deliberative ones? In Proceedings of the IJCAI, 1993.

No context found.

Knight K.: Are Many Reactive Agents Better Than a Few Deliberative Ones? In: Proc. IJCAI '93, Chambery, 1993, 132--137

No context found.

Knight, K.: Are Many Reactive Agents Better Than a Few Deliberative Ones? In: Proc. IJCAI '93, Chambery, 1993, pp. 132--137.

No context found.

Knight, K.: Are Many Reactive Agents Better Than a Few Deliberative Ones? In: Proc. IJCAI '93, Chambery, 1993, pp. 132--137.

No context found.

K. Knight, "Are Many Reactive Agents Better than a Few Deliberative Ones?" IJCAI-93, pp. 432#437, 1993.

No context found.

Knight K.: Are Many Reactive Agents Better Than a Few Deliberative Ones? In: Proc. IJCAI '93, Chambery, 1993, 132--137

No context found.

Knight, K. 1993. Are many reactive agents better than a few deliberative ones? In Proc. 13th International Joint Conference on Artificial Intelligence (IJCAI-93), 432--437.

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