GDSS and the structure of decisions

Yesterday, Simone dealt with Group Decision Support Systems (GDSS) in her blog entry, in which she introduced four different types of GDSS-Systems based upon the characteristics of the group.

However, GDSS-Systems do not only need to adapt to the group, they also need to be chosen according to the types of decisions that need to be made. In general, GDSS do not make the decisions themselves, they rather present information that are necessary to make a decision.

According to scientist Mallach there are three components of decisions, which decide about the nature of the decision process. They apply to decisions of individuals as well as to group decisions.

The first component is the data that is collected by a decision maker and that will by used in making the decision. The more data there is the more information can be included in the decision process. The second component is the process selected by the decision maker to combine the data. Last but not least, the evaluation and learning component compares the decisions and examines them in order to see if there is a need either to change the data or the process that combines the data.

According to those components three categories of decisions can be made considering the degree of structure in the decision-making activity.

1. Structured decisions are usually preplanned and they are made for routine tasks. Usually, those decisions have been made before. All the three components can be well specified: the data, process and evaluation. They are routine and predictable choices.

For this kind of decisions it is possible to develop computer programs that collect the data and combine it. But since the process is persistent and logical, a GDSS is usually not necessary. Instead, it makes more sense to use an automated tool, which can make the decision based on the predefined criteria.

2. The phases of decision making process of unstructured decisions are on the contrary irregular although they have the same components. However, there is little agreement on their nature. The problem is, that each decision maker uses different data and processes and therefore the reached conclusions can be totally different from each other.

For this kind of decisions programs begin by focusing on the team are needed, for example by simulating the process used by the group. They have been called “expert systems” because they are based upon the group’s expertise or experience. However, an expert decision maker won’t likely be replaced by such a system.

Another approach is to monitor and document the process that was used. This way, the decision makers can review what has been done so far at any time.

3. Most of the decision support systems focus on semi-structured decisions. In general, they have some structured aspect, for example they have some agreement on the data, process and/or evaluation. In order to find out which support system is needed, it is necessary to detect the borders of the decision maker. For example, if the data is incomplete the system will provide all the essential data.

http://en.wikibooks.org/wiki/Systems_Theory/Decision_Structure

http://www.referenceforbusiness.com/encyclopedia/Cos-Des/Decision-Support-Systems.html