| S. Prestwich. Combining the scalability of local search with the pruning techniques of systematic search. Annals of Operations Research, 115:51--72, 2002. |
.... of local search unnaturally, but empirical evidence supports our claim: an IDB algorithm was shown in [21] to scale almost precisely like local search (Walksat) on random 3 SAT problems, and another IDB algorithm scales like the Min Conflicts Hill Climbing algorithm on the N queens problem [19]. Moreover, there exist a few other local search algorithms that explore partial solutions, for example the IMPASSE graph colouring algorithm [14, 17] 3 Experimental results As a test bed we use a set of benchmarks from the Computational Symposium on Graph Coloring and its Generalizations held ....
....[3] uses a limited form of backtracking (closer to chronological backtracking than to IDB) MAC IDB replaces chronological by incomplete dynamic backtracking, and forward checking by arc consistency. A variety of other constraint programming local search hybrids have been devised and surveyed in [6, 19], but these do not generally maintain arc or other consistency within local search. For example, systematic backtracking has been used to explore local search neighbourhoods [23] conversely, local search has been used within a complete backtracker to select the best branching variable [2] ....
S. D. Prestwich. Combining the Scalability of Local Search with the Pruning Techniques of Systematic Search. Annals of Operations Research, to appear.
....as unassigned ones. In the example of Figure 4 2 the variables for rows 2 5 each have domain size 1, but in general these may be different. This can be used to guide the selection of variables for unassignment, for example the variable with greatest domain size. These techniques are explored in [9], where such an algorithm is shown to require even fewer search steps than Min Conflicts Hill Climbing to solve N queens problems. Because this approach performs local search in a constrained space, we call it Constrained Local Search (CLS) Alternatively it can be viewed as a non systematic form ....
....local search in a constrained space, we call it Constrained Local Search (CLS) Alternatively it can be viewed as a non systematic form of backtracking, which we call Incomplete Dynamic Backtracking. The relationships between CLS and other search algorithms are discussed in previous papers [14, 10, 9]. In the next section we describe a CLS implementation for BIBD generation. 2 Local search for BIBD generation BIBD generation is a standard combinatorial problem, originally used in the statistical design of experiments but since finding other applications such as cryptography. A BIBD is ....
S. Prestwich. Combining the Scalability of Local Search with the Pruning Techniques of Systematic Search. Annals of Operations Research (to appear).
No context found.
S. Prestwich. Combining the scalability of local search with the pruning techniques of systematic search. Annals of Operations Research, 115:51--72, 2002.
No context found.
Prestwich, S.: Combining the scalability of local search with the pruning techniques of systematic search. Annals of Operations Research 115 (2002) 51--72
No context found.
Prestwich, S.: Combining the scalability of local search with the pruning techniques of systematic search. Annals of Operations Research 115 (2002) 51--72
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