The proposed algorithms use a best first search technique and report the solutions using an implicit representation ordered by cost. In this paper, we present two versions of the search algorithm -- (a) an initial version of the best first search algorithm, ASG, which may present one solution more than once while generating the ordered solutions, and (b) another version, LASG, which avoids the construction of the duplicate solutions. The actual solutions can be reconstructed quickly from the implicit compact representation used. We have applied the methods on a few test domains, some of them are synthetic while the others are based on well known problems including the search space of the 5-peg Tower of Hanoi problem, the matrix-chain multiplication problem and the problem of finding secondary structure of RNA. Experimental results show the efficacy of the proposed algorithms over the existing approach. Our proposed algorithms have potential use in various domains ranging from knowledge based frameworks to service composition, where the AND/OR structure is widely used for representing problems.
Logic, and declarative representation of knowledge in general, have long been a preferred framework for problem solving in AI. However, specific subareas of AI have been eager to abandon general-purpose knowledge representation in favor of methods that seem to address their computational core problems better. In planning, for example, state-space search has in the last several years been preferred to logic-based methods such as SAT. In our recent work, we have demonstrated that the observed performance differences between SAT and specialized state-space search methods largely go back to the difference between a blind (or at least planning-agnostic) and a planning-specific search method. If SAT search methods are given even simple heuristics which make the search goal-directed, the efficiency differences disappear.
Wheeler Ruml Division of Engineering and Applied Sciences Harvard University 33 Oxford Street Cambridge, MA 02138 ruml@eecs, harvard, edu Abstract When searching a tree to find the best leaf, complete search methods such as depth-first search and depth-boundediscrepancy search use a fixed deterministic order that may or may not be appropriate for the tree at hand. Adaptive probing is a recently-proposed stochastic method that attempts to adjust its sampling online to focus on areas of the tree that seem to contain good solutions. While effective on a variety of trees, adaptive probing wastes time learning basic features of the problem that are built into other algorithms, such as the fact that the heuristic is often helpful. In this paper, we investigate two simple methods for adding such prior knowledge to adaptive probing. The second uses a heuristically biased policy at the start of the search, gradually deferring to learned information later iterations.
Next time you see a cute pair of shoes or a cool shirt you'd like to buy, snap a picture or take a screenshot. "Find It On eBay" gives you the power to share images straight from any social network or website to the online shopping platform's application. Just choose the website's logo with the "Find it now" tag line, click "search using this image" when it pops up and highlight the part of the photo you want to look up. The other feature that's simply called "Image Search" gives you the power to look for items using photos you've taken or saved on your device. Both tools make it much easier to find listings when you're looking for something really specific or looking up something you have no idea how to describe -- hey, it happens to everyone.
We consider the challenging practical problem of optimizing the power production of a complex of hydroelectric power plants, which involves control over three continuous action variables, uncertainty in the amount of water inflows and a variety of constraints that need to be satisfied. We propose a policy-search-based approach coupled with predictive modelling to address this problem. This approach has some key advantages compared to other alternatives, such as dynamic programming: the policy representation and search algorithm can conveniently incorporate domain knowledge; the resulting policies are easy to interpret, and the algorithm is naturally parallelizable. Our algorithm obtains a policy which outperforms the solution found by dynamic programming both quantitatively and qualitatively. Papers published at the Neural Information Processing Systems Conference.