this book proposes an approach to analyzing, designing and implementing group decision support systems (gdss). the experience gained in developing a concrete system, co-op, suggests that gdss should be distributed, loosely-coupled and process-driven. a distributed and loosely-coupled gdss architecture provides autonomy and flexibility for individual decision-making, and homogeneity and simplicity for group problem solving. also, process-driven gdss are able to deal with the unpredictable nature of group problems since collective decision processes have been shown to be the only elements in a gdss that are (i) stable enough to fit into most collective problems, (ii) reasonably structurable to be implementable, and (iii) sufficiently controllable to guarantee appropriate use. from a multiple criteria decision-making (mcdm) viewpoint, this book supports the integrated use of various mcdm methods to help a gdss (i) support a wide range of decision situations, (ii) attenuate the difficulty of information search, (iii) allow division of decision-making tasks, and (iv) permit consensus seeking analysis. co-op runs on a network of individual workstations. it contains a set of mcdm methods, techniques of aggregation of preferences, and a consensus seeking algorithm to support negotiation. electronic communication among group members is monitored by a group norm filter which is adaptable to a large number of collective decision situations. this book also reports some empirical evaluations of co-op and expands the proposed approach to non-cooperation and organizational decision-making.
