## Sweep A*: Space-efficient heuristic search in partially ordered graphs (2003)

### Cached

### Download Links

- [www.cs.iastate.edu]
- [www.cs.msstate.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | In Proceedings of the 15th IEEE International Conference on Tools with Artificial Intelligence |

Citations: | 17 - 4 self |

### BibTeX

@INPROCEEDINGS{Zhou03sweepa*:,

author = {Rong Zhou and Eric A. Hansen},

title = {Sweep A*: Space-efficient heuristic search in partially ordered graphs},

booktitle = {In Proceedings of the 15th IEEE International Conference on Tools with Artificial Intelligence},

year = {2003},

pages = {427--434}

}

### Years of Citing Articles

### OpenURL

### Abstract

We describe a novel heuristic search algorithm, called Sweep A*, that exploits the regular structure of partially ordered graphs to substantially reduce the memory requirements of search. We show that it outperforms previous search algorithms in optimally aligning multiple protein or DNA sequences, an important problem in bioinformatics. Sweep A * also promises to be effective for other search problems with similar structure. 1.

### Citations

396 |
Real-time heuristic search
- Korf
- 1990
(Show Context)
Citation Context .... The approach we describe can be used for any search problem, and not only sequence alignment. We propose a variation of realtime A* (RTA*), a search algorithm that creates a solution path in stages =-=[6]-=-. We modify the algorithm so that it uses all available memory, but never exceeds it. Our version of RTA* searches forward from the current state until a memory limit is reached. Then it selects the n... |

270 |
A linear space algorithm for computing maximal common subsequences
- Hirschberg
- 1975
(Show Context)
Citation Context ...her path, in order to prevent duplicate search effort. Inspired by Hirschberg’s divide-andconquer technique for reducing the memory requirements of dynamic programming for multiple sequence alignment =-=[2, 12]-=-, Korf and Zhang [9] introduced a variant of A* called Divide-and-Conquer Frontier A* (DCFA*) that prevents duplicate search effort without storing a Closed list; only an Open list is used. DCFA* has ... |

211 |
Depth-first iterative deepening : an optimal admissible tree search
- Korf
- 1985
(Show Context)
Citation Context ... search frontier, and a Closed list to store already expanded nodes. The space needed to store all explored nodes can quickly fill available memory. Although limitedmemory variants of A* such as IDA* =-=[5]-=- and RBFS [7] conserve memory by not storing Open and Closed lists, they are ineffective for complex graph-search problems with many duplicate paths, such as the multiple sequence alignment problem, d... |

180 | Optimal Alignments in Linear Space
- Myers, Miller
- 1988
(Show Context)
Citation Context ...her path, in order to prevent duplicate search effort. Inspired by Hirschberg’s divide-andconquer technique for reducing the memory requirements of dynamic programming for multiple sequence alignment =-=[2, 12]-=-, Korf and Zhang [9] introduced a variant of A* called Divide-and-Conquer Frontier A* (DCFA*) that prevents duplicate search effort without storing a Closed list; only an Open list is used. DCFA* has ... |

155 |
The Multiple Sequence Alignment Problem in Biology
- Carillo, Lipman
- 1988
(Show Context)
Citation Context ...own that the multiple sequence alignment problem can be formulated as a shortest-path problem in a graph that corresponds to an n-dimensional lattice, where n is the number of sequences to be aligned =-=[1]-=-. Figure 2 shows an example of a two-dimensional lattice for a pairwise alignment problem, and Figure 1 shows an example of a three-dimensional lattice corresponding to a problem of aligning three seq... |

145 |
Linear-space best-first search
- Korf
- 1993
(Show Context)
Citation Context ...ier, and a Closed list to store already expanded nodes. The space needed to store all explored nodes can quickly fill available memory. Although limitedmemory variants of A* such as IDA* [5] and RBFS =-=[7]-=- conserve memory by not storing Open and Closed lists, they are ineffective for complex graph-search problems with many duplicate paths, such as the multiple sequence alignment problem, due to excessi... |

48 | Divide-and-conquer frontier search applied to optimal sequence alignment
- Korf, Zhang
- 2000
(Show Context)
Citation Context ...tperforms dynamic programming in finding optimal alignments by using a lower-bound function (i.e., an admissible heuristic) to direct the search and limit how much of the state space must be explored =-=[4, 10, 14, 9, 11, 3, 16]-=-. It is well known that the limiting factor in scaling up A* is its memory requirement. A* keeps track of the explored part of the search space using an Open list to store nodes on the search frontier... |

25 |
Divide-and-conquer bidirectional search: First results
- Korf
(Show Context)
Citation Context ...lled Sweep A*, that exploits the structure of partially ordered graphs to substantially reduce the memory requirements of search. Like other variants of A* that do not keep all closed nodes in memory =-=[8, 9, 16]-=-, Sweep A* has two stages: first it searches forward from the start node to the goal node in order to find the optimal cost of a solution, as well as one or more intermediate nodes along an optimal pa... |

24 |
A* with partial expansion for large branching factor problems
- Yoshizumi, Miura, et al.
- 2000
(Show Context)
Citation Context ...tperforms dynamic programming in finding optimal alignments by using a lower-bound function (i.e., an admissible heuristic) to direct the search and limit how much of the state space must be explored =-=[4, 10, 14, 9, 11, 3, 16]-=-. It is well known that the limiting factor in scaling up A* is its memory requirement. A* keeps track of the explored part of the search space using an Open list to store nodes on the search frontier... |

19 | Sparse-memory graph search
- Zhou, Hansen
- 2003
(Show Context)
Citation Context ...tperforms dynamic programming in finding optimal alignments by using a lower-bound function (i.e., an admissible heuristic) to direct the search and limit how much of the state space must be explored =-=[4, 10, 14, 9, 11, 3, 16]-=-. It is well known that the limiting factor in scaling up A* is its memory requirement. A* keeps track of the explored part of the search space using an Open list to store nodes on the search frontier... |

18 | The Practical Use of the A Algorithm for Exact Multiple Sequence Alignment - Lermen, Reinert - 2000 |

16 |
Enhanced A* algorithms for multiple alignments: optimal alignments for several sequences and k-opt approximate alignments for large cases
- Ikeda, Imai
- 1999
(Show Context)
Citation Context |

15 | Memory-efficient A* heuristics for multiple sequence alignment
- McNaughton, Lu, et al.
- 2002
(Show Context)
Citation Context |

11 |
Multiple sequence alignment using anytime a
- Zhou, Hansen
(Show Context)
Citation Context ... lower-bound estimate (or f-cost) of any node is greater than the upper bound, the node is not inserted in the Open list because it cannot lead to an optimal solution and will never be expanded by A* =-=[4, 15]-=-. Both techniques have proved effective in improving the scalability of A* for multiple sequence alignment. Other researchers have developed techniques for reducing the size of the Closed list. The Cl... |

10 |
Comparing best-first search and dynamic programming for optimal multiple sequence alignment
- Hohwald, Thayer, et al.
- 2003
(Show Context)
Citation Context |

4 |
Searching graphs with A*: Applications to job sequencing
- Sen
- 1996
(Show Context)
Citation Context ...aving constant ∆ is 1. 2.2.3. Job sequencing The problem of job sequencing is to schedule N jobs sequentially on a machine such that a penalty function on job completion time is minimized. Sen et al. =-=[13]-=- show that A* is effective in solving this problem. Figure 3(b) shows the partially ordered state-space graphsFigure 3. (a) shows a partially ordered state-space graph for determining the most probabl... |