Wed. Nov. 20, 10:30 to 12:00
Room C104, Colorado Conference Center, Denver CO
1. Christopher Ellis, Douglas M. Johnston and Lewis D. Hopkins
University of Illinois - Urbana/Champaign
Ethnographic Assessment of Ecological Modelers for Design of a Collaborative Geographical Modeling System
2. David S. Lemberg, Mike Figueroa and Richard L. Church
University of California
Feasible Alternatives Generation in Collaborative Spatial Decision Making
3. Piotr Jankowski
University of Idaho
Architectures for Space and Time Distributed Collaborative Spatial Decision Making
4. Marc P. Armstrong and Richard J. Marciano
The University of Iowa and San Diego Supercomputer Center
Distributed Parallelism: Impacts on GIS and Collaborative Spatial Decision-Making
This paper describes the results of an ethnographic study of the work practices of land managers and ecological modelers. Results were used to design a geographical ecological modeling system which allows land managers and ecological modelers to work asynchronously in a common environment, drawing and building on the work and information (models, data, etc.) of each other. These systems should support the work users conduct individually while facilitating communication with other users. Ethnographic methods such as interviews (Hutchins, Bucciarelli, Randall), identification of boundary objects (Seely-Brown, Star), and task analysis (Rasmussen, Gasser) help uncover practices that are deeply rooted in the work of diverse system users while simultaneously helping identify shared resources and issues that are common to other system users.
Initial analysis focused on ecological modelers. Different modelers approach modeling at different scales from individual species models to landscape level models. This suggests that a spatially hierarchical system design may best take advantage of these work approaches. Information which planners and modelers receive from others often comes in the form of artifacts (data, models, or visualizations in portable form) made available at some other time and/or place. The system helps planners and modelers make decisions about the types and spatial and temporal scales of data needed and leads to standardization reducing confusion in communication between different groups. Public memory space is used to post and retrieve results from each level of work. Public libraries of analytical representation routines in conjunction with private modeling space provide resources supporting the modeling task.
Collaborative Spatial Decision Making (CSDM) is the process of groups of decision makers and interested parties using Geographic Information Analysis tools and Group Decision Support tools to generate or negotiate better solutions to public problems. While there are many applications of spatial modeling and GIS in the public decision making process, there are few applications that focus on cooperative problem formulation such that the models can generate solutions that are both optimal and politically feasible. A group may create a feasible model by cooperatively agreeing on a list of objectives, decision variables, constraints, and weighting factors.
This lecture with slides will demonstrate the conception of a collaborative problem formulation and alternative generation system for the School District Planning Problem (SDPP). A group working on the SDPP seeks to create a long-term student enrollment management plan for a school district. Major interests include the students, parents, neighbors, teachers, staff, administrators, and Board Members (decision makers and political actors). Major obstacles include changing demographics, limited funds, limited space, existing rules and regulations, political considerations, and interest group activism. Some general (and conflicting) goals include maximizing education quality, student safety, neighborhood stability, equity, and diversity, while minimizing costs. We propose a GIS tool where groups may select the objectives and constraints to develop feasible solutions specific to their own school district's needs. The group can specify boundary conditions, site capacities and locations, student travel time and distance constraints, etc., to generate modeled alternatives which may be evaluated and adjusted to feasible options.
The increasing involvement of stakeholder groups in solving spatial decision problems have created a need for information technology capable of supporting collaborative spatial decision making (CSDM). Such information technology has developed in recent years for the computerized support of group decision making aimed at solving business problems. Similar information technology is now being developed to support group decision making aimed at solving spatial decision problems, e.g., site selection, choice of environmental and economic strategies, and urban/regional development. However, the research on benefits (or results) of such technology has been limited so far to meeting situations (same place, same time). This paper presents early results of work on supporing collaborative spatial decision making in distributed space and time environment, i.e., people working at different locations and at different times. Of particular interest in this work are different architectural arrangements of decision support tools including maps, multiple criteria decision models, and communication tools. The paper examines the implementation costs, efficiency, and utility of internet and intranet architectures supporting CSDM.
Richard J. Marciano
Computational Environmental Sciences Group
San Diego Supercomputer Center
P.O. Box 85608
San Diego, CA 92186-9784
When groups of people come together to explore the geographical dimensions of public policy issues, scheduled meetings have a finite (and often short) duration. If a GIS-based model requires a substantial amount of time to execute, the number of alternatives that can be generated and evaluated during each meeting is diminished. In this paper we present results from research that evaluates alternative approaches to computationally-complex GIS-based analyses, especially as they relate to improving the performance of distributed, networked, collaborative decision-making environments. The first element of the paper centers on the use of a network of workstations (NOW) to support parallel interpolation for a large problem (1024 x 1024 cells with 10,000 control points). Based on a division of the interpolation algorithm into a set of mutually independent processes, computational experiments using NOW achieves a level of performance that far exceeds that of the current generation of workstations. In fact, the results compare favorably with those obtained from a dedicated parallel supercomputer (Cray T3D). The experiments also show general promise for a large class of (grid cell) GIS algorithms that can be divided into large (e.g. subroutine or procedure) sub-problems. These results are especially significant given that workstation-based computing now dominates GIS. These workstations, which are also the predominant technology used in collaborative decision-support environments present a ready source of spare computing power that can be accessed and used during the period of a typical group meeting.