Searching spatial configurations is a particular case of maximal constraint satisfaction prob-lems, where constraints expressed by spatial and non-spatial properties guide the search process. In the spatial domain, binary spatial relations are typically used for specifying constraints while searching spatial configurations. Searching configurations is particularly intractable when con-figurations are derived from a combination of objects, which involves a hard combinatorial problem. This paper presents a genetic algorithm that combines a direct and an indirect ap-proach to treating binary constraints in genetic operators. A new genetic operator combines randomness and heuristics for guiding the reproduction of new individuals in a population. In-dividuals are composed of spatial objects whose relationships are indexed by a content measure. This paper describes the genetic algorithm and presents experimental results that compare the genetic versus a deterministic and a local-search algorithm. These experiments show the con-venience of using a genetic algorithm when the complexity of the queries and databases do no guarantee the tractability of a deterministic strategy. Index Terms: Evolutionary computation, genetic algorithm, geographic information systems, constraint satisfaction problems, information retrieval.

Abstract. A constant challenge of current spatial information systems is the retrieval of spatial configurations. This paper describes a new approach to retrieving spatial information whose novelty lies in using a content measure of topological relations to search and index spatial configurations. This approach uses a tree-based schema to index relations between objects’ minimum bounding rectangles, it preprocesses the user query to obtain an ordered list of spatial constraints, and it explores three searching algorithms that work over an indexed domain: full-restrictive forward checking, partial-restrictive forward checking, and permutation-based searching. Experimental results show the viability of this type of indexing schema as well as the differences among searching algorithms. 1

degree at The University of Maine, I agree that the Library shall make it freely available for inspection. I further agree that permission for "fair use " copying of this thesis for scholarly purposes may be granted by the Librarian. It is understood that any copying or publication of this thesis for financial gain shall not be allowed without my written permission.

Hand drawn sketches can be used as an aid in spatial queries for identifying spatial relationships. This paper is developing a framework describing the computational methods for solving these queries when a spatial reference data set is given. It proposes a suitable representation of the spatial data and how constraints in the data set can be modeled for the purpose of sketch interpretation. Topologic relationships between spatial objects are the most important source of information which is applied in this context. The discrete relaxation algorithm is used to find a set of objects which is consistent with the constraints from the relationship graph. Finally a sketchpad application, developed for experiments with the algorithm, is presented. 1.