Showing papers by "Indian Institute of Management Calcutta published in 1989"
Proceedings Article•
20 Aug 1989TL;DR: Extensive runs on a variety of search problems, involving search graphs that may or may not be trees, indicate that MREC with M = 0 is as good as IDA* on problems such as the 15- puzzle for which IDA*, is suitable, while M REC with large M is as fast as A*, on problems for which node expansion time is not negligible.
Abstract: MREC is a new recursive best-first search algorithm which combines the good features of A* and IDA*. It is closer in operation to IDA*, and does not use an OPEN list. In order to execute, all MREC needs is sufficient memory for its implicit stack. But it can also be fed at runtime a parameter M which tells it how much additional memory is available for use. In this extra memory, MREC stores as much as possible of the explicit graph. When M = 0, MREC is identical to IDA*. But when M > 0, it can make far fewer node expansions than IDA*. This can be advantageous for problems where the time to expand a node is significant. Extensive runs on a variety of search problems, involving search graphs that may or may not be trees, indicate that MREC with M = 0 is as good as IDA* on problems such as the 15- puzzle for which IDA* is suitable, while MREC with large M is as fast as A* on problems for which node expansion time is not negligible.
85 citations
Proceedings Article•
20 Aug 1989TL;DR: It is shown that an earlier algorithm of the authors, called ITERSSS*, which allows SSS* to run in restricted memory, can also be parallelized using the above scheme.
Abstract: PARSSS* is a parallel formulation of SSS* that is suitable for shared-memory multiprocessor systems. It is based on the distributed tree search paradigm of Ferguson and Korf. The main difficulty in parallelizing SSS* lies in achieving proper coordination between processes running on different subtrees of the game tree. This has been resolved in PARSSS* by the use of a shared array which maintains summary information on all processes that are currently in execution. Problem-independent speed-up values for PARSSS* have been obtained experimentally. It is shown that an earlier algorithm of the authors, called ITERSSS*, which allows SSS* to run in restricted memory, can also be parallelized using the above scheme.
2 citations
TL;DR: In this paper, the authors studied the weak convergence of the residual median process when the observations follow a strictly stationary o-mixing process and when hte observations are randomly censored.
Abstract: This article studies the weak convergence of the residual median process (i) when the observations follow a strictly stationary o-mixing process and (ii) when hte observations are randomly censored. In both these cased the residual median prodeas converges weakly to a gaussian process.
1 citations
11 Dec 1989
TL;DR: A new arc-marking algorithm, namely MarkC is presented, which uses the run-time heuristic modification scheme, and the node selection criterion of AlgorithmC, which has some important properties under general inadmissible heuristics.
Abstract: A new arc-marking algorithm, namely MarkC is presented MarkC uses the run-time heuristic modification scheme (due to Mero), and the node selection criterion of AlgorithmC (due to Bagchi and Mahanti) AlgorithmC, based on the node expansion scheme, has some important properties under general inadmissible heuristics AlgorithmC always finds a least costly solution path of V (an implicitly defined set of nodes) and makes O(N2) node expansions at the worst It is shown that MarkC retains all the merits of C, yet it does not iterate more number of times than C But there are examples where C can make some more iterations than MarkC Finally, a comparative study is presented by summarizing the results on Algorithms MarkA, C, and MarkC