A Markov chain analysis of a computer supported information network system-One approach to a model of building consensus in electronic brainstorming

The study of Group Decision Support Systems (GDSS) has matured considerably over the past five or six years. It is suggested that the chief value of GDSS is that they help improve communication. This paper reports on a Markov and a combinatorial model of a GDSS for idea generation, and particularly focuses upon issues on building consensus in Electronic Brainstorming (EBS) session of GDSS. The present model can explain one aspect of dynamics of an EBS and an idea generation process. Group members of GDSS are located in separated places. Each group member holds a set of different information at the beginning of the task. These group members communicate among themselves by a computer supported information network system. Three communication types are categorized according to the transmission mode, transaction pattern, number of transaction pairs and the transaction direction. The paper compares three communication types A, B and C based on the performance measured by the wait time for consensus and the cost measured by the complexity of protocols and the number of message buffers in each node of the network. When the communicaton type in EBS is selected, the trade-off among three communication types must be considered. Type A is the lowest one of all based on the performance measured by the wait time but is the best one of all based on the cost measurement. Type B is the highest one of all based on the performance measured by the wait time but is the worst one of all based on the cost measurement. Type C is the middle of all. For designers who are interested in developing their own GDSS facilities, these results are useful.