Optimization in Geodesign

Abstract

The use of optimization algorithms in landscape decision-making has been seen as the antithesis of a design process. Indeed, Michael F. Goodchild clearly distinguished optimization, small-d design as part of operations research and management science, from big-d Design, which is the province of landscape architecture and planning[1]. This paper argues a case for integration of optimization techniques and geodesign technologies to help shape, but not narrow, the decision space associated with a landscape design or planning situation. Optimization need not, as Goodchild argues, be limited to production of a single point in the solution space. By creative manipulation of objectives and constraints (both spatial and global) a disparate set of possible futures can be generated. The advantage of using optimization is that all these points within the solution space are on a non-dominated surface. When the user is limited to a set of interactive geodesign tools, sketching shapes and editing attributes, many of the generated solutions may be inferior, on all criteria, to another solution. Optimization removes these inferior solutions and makes it easier for the designer, and the stakeholders, to review the options and make decisions. The interactive processes of spatial selection and attribution are still important; the challenge is to build systems that link these capabilities to robust optimization algorithms. The paper reviews these possibilities in the context of forest management and regional landscape planning and explores options for achieving fast response times, which supporting selection among the non-inferior solutions and dealing with uncertainty and conditions that change over time.

Keywords

Geodesign / Optimization / Forest Management / Regional Planning

RIGHTS & PERMISSIONS

2014 Higher Education Press