Decision Support Systems - Overview

A Decision Support System or DSS is defined for this directory as:

"a model or framework that can be used to assist with analysis

and decision-making processes within an organisation".

This broad definition is used to allow the inclusion of a wide range of models, frameworks or tools as long as they have been used or are being developed for a purpose that is useful to Councils decision-making needs. 

A DSS can be as simple as a spread sheet with some linked formula through to a complex integration of several spatially applied models.

Types of Decision Support Systems

A wide range of DSSs is included in this directory and they have been divided into management domains to allow for searching the directory.  The terminology used in this directory recognises that there can be different levels and complexity of DSS's.  This is illustrated in Figure 1.

At the simplest level there are the key assessment methods that can underpin a DSS (e.g. Multi Criteria Analysis, Bayesian Belief Network) or a framework (i.e. deliberation process). In some cases these can be used directly by end users to help them make decisions or solve problems. Often, however, these assessment methods are  incorporated into a computer based model to form a specific DSS product. These computer models can be either spatial or non-spatial. They can range in complexity from a single model based on one assessment method (e.g. Overseer - Input/Output model) to several interlinked models within an integrated DSS that incorporates several assessment methods (e.g WISE - Input/Output, Spatial dynamic, GIS).

 Types of DSSs3

 Figure 1: Types of Decision Support Systems

 

Assessment Methods/Frameworks

Assessment methods or frameworks are the underlying methods or frameworks which is used to manipulate, assess or model data. These methods can vary from simple processes to very complex frameworks. At the simplest level an assessment method can be used as a decision support system or it can be incorporated in a system that can include one or more methods. Councils have already been using methods such as Cost Benefit Analysis, Environmental Impact Assessments, Risk Assessment, and Scenario Planning.  There is a wider range of assessment methods available, many of which are now incorporated into computer models as decision support systems.

Assessment methods have been grouped into several categories by van den Belt et al, 2010:

  • Input-Output Modelling - Input-output analysis provides a comprehensive snapshot of the structure of the inter-industry linkages in an economy. An input-output model may be used to trace the direct, indirect and induced economic impacts associated with a given change in final demand.
  • Geographical Information Systems (GIS) - GIS modelling presents data in map form as well as providing a tool for map-based queries and analyses. Numerical data layers can describe aspects such as climate, landforms, income levels, soils, asset, business locations, and so on.
  • Spatially Dynamic Support Systems (SDSS) - Spatially dynamic systems have their roots in GIS's. They apply a variety of techniques to simulate the dynamics of land-use change at various spatial scales. The desire for such modelling has been driven by the need to balance environmental, social, and economic consideration in decision-making processes.
  • Mediated Modelling - Mediated Modelling is a participatory approach that uses computer modelling as a consensus building tool. This model builds in social decision-making processes with advances in computing capabilities to provide facilitated, computer-assisted processes in group settings.
  • General Equilibrium Modelling - Computer based General Equilibrium  models have become invaluable tools for analysing the economic impacts of environmental policies and environmental impacts of economic policies. GE models provide a comprehensive and detailed description of an economy that is based on microeconomic foundations and is consistent with key macroeconomic balances and principles. They may be readily extended to model resource use, emissions and other environmental pressures that are directly associated with production or consumption activities.
  • Multi-Criteria Analysis (MCA) - MCA techniques can be used to support choices within a set of defined options, particularly when decision-makers are concerned with multiple dimensions of performance that are  not directly commensurable. More recent use of MCA techniques has focused on multiple decision-makers with different values, and where uncertain and/or subjective aspects of performance are important.
  • Agent Based Modelling (ABM) - ABM reflects the likely or expected behaviour of various agents or stakeholders in a system. ABM is often spatially explicit and can be used in conjunction with role-play situations to test a model interactively in a real life context. ABM is widely used internationally in Land Use-Land Change studies.
  • Bayesian Belief Networks (BBN) - A BBN is a probabilistic graphical model (a type of statistical model) that represents a set of random variables and their conditional dependencies. The model structure allows for the representation of and reasoning about an uncertain situation. The nodes in a network represent a set of variables in the domain being modelled. The nodes are connected by links representing the relationship between variables.

 

Spatial DSSs

Spatial DSSs have been designed to support decision-making processes for complex spatial problems. They have become more available and advanced with the development of computer power, and geographical information systems and the creation of spatial database to support the processes. Because of their increased complexity many spatial DSSs have high learning thresholds for use and can require specialist support and advice to implement.

Spatial DSS's are required to have additional capabilities and functions such as:

  • mechanisims for input of spatial data,
  • representation of the complex spatial relations and structures that are common in spatial data,
  • analytical techniques that are unique to both spatial and geographic analysis, and
  • outputs in a variety of spatial forms including maps and other more specalised types.

The characteristics of spatial DSSs facilitate a decision process that can be more iterative, integrative and participative.

Non-Spatial DSSs

Non-Spatial DSSs have been in use much longer than the Spatial DSS. As the name implies they are used to support decision making that is based on an issue or problems that is not spatially assessed within the tool - although the decision may be applied spatially. Non-spatial DSS' generally tend to utilise one analytical technique and can be computer based. They can also be incorporated into more complex computer models to become part of a spatial DSS.

 

Roles of DSSs in Policy development and Implementation

Policy development and implementation involves a range of decision needs through the Plan-Do-Check-Review cycle (Fig. 2).

This cycle identifies four main processes:agenda setting - identifies the issues and defines policy objectives that define the expected outcomes; policy formulation - defines and analyses the range of policy instruments that could be applied to achieve the objectives and provides a set of policy methods (i.e. rules, incentives); policy implementation - takes these methods and allocates resources to applying them; policy evaluation - the final process in the cycle monitors the results of implementing the methods and evaluates the results against anticipated results of the policy.

Opportunities exist to use DSS's to help issue identification as part of agenda setting, provide information and support for community engagement processes as well as identifying, analysing, and evaluating policy options.  DSS's can also be used to support management decisions as part of implementation activities and to assist in the assessment of monitoring of outcomes when evaluation of policy is undertaken. 

 

Policy Process generic

 

Figure 2: Generic Policy Process 

 

Selecting a DSS for your needs - here are some tips