Exploring the Solution Space of Semi- Structured Spatial Problems Using Genetic Algorithms

The resolution of semi-structured spatial problems often requires consensus building and compromise among stakeholders as they attempt to optimize their own set of criteria. The union of these sets form a criteria space that constrains the set of viable solutions that may be adopted by decision-makers. Knowledge about the criteria space, the solution space, and the relation between the two is normally incomplete and this lack of understanding places real limits on the ability of decision-makers to solve complex spatial problems. This research explores new approaches that are designed to establish a link between criteria space and solution space and to provide a mechanism that competing stakeholders can use to identify areas of conflict and compromise. 1.0 INTRODUCTION Spatial problem solving often requires collaboration among multiple decision-makers because the effects of spatial decisions often cut across traditional bounds of discipline, jurisdiction, and ownership. Because different decision-makers will have different views of a problem, the evaluation of alternative solutions to it is complicated since: 1) a collection of spatial models and analytical tools is needed to evaluate how well each alternative meets stated criteria; 2) multicriteria evaluation tools are needed to integrate the results of these models and tools; 3) the set of all possible solutions (the solution space) is often intractable (theoretically infinite for field-based problems); and 4) not all criteria are well articulated or even known at the beginning of an analysis (i.e., spatial problems are often semi-structured). Furthermore, the resolution of semi-structured spatial problems often requires consensus building and compromise among decision-makers because as individuals attempt to optimize their own set of criteria they will often come into conflict with others. The