The network topology design problem has long been recognized as extremely complex and very quickly becomes nmanageable as the size of the network increases. Existing heuristic design procedures are quite efficient for the design of small networks (10-25 nodes); however, they become very costly and even prohibitive when dealing with large tworks. The routing and scheduling of vehicles is an area of practical importance in computer and transportation management. There are many variations of the vehicle routing problem ranging from school bus routing to the dispatching of delivery trucks for consumer goods. In all cases, the basic components of the problem are a fleet of vehicles with fixed capabilities (capacity, speed, etc.) and a set of demands for transporting certain objects (consumer goods, school children.between specified pickup and delivery points. Th routing decision involves determining which of th demands will be satisfied by each vehicle and wha route each vehicle will follow in servicing it assigned demand. These decisions should be made t minimize the cost of operating the vehicle fleet. Vehicle routes must also satisfy a variety oz constraints arising from factors such as fixed vehicle capacity and union regulations on driver work schedules. Three methods of solution are considered in this report: (A) The sweep approach. (B) The "savings" method. (C) The clustering approach. A design methodology based on the clustering of the network nodes is presented in this report in order to alleviate the computational cost and computer storage involved in the design. More specifically, the emphasis is on the determination of an adaptative clustering technique which is used to solve the vehicle-dispatching problem. In terms of solution time, the clustering approach outperforms the sweep and "savings" heuristics, but in terms of quality of the results, the closest competitor is the Gillet and Miller rithod. In addition, the clustering heuristic can be improved by using efficient data structures. The computational results are illustrated with a routing study for a set of standard data bases. The general methodology (based on the fuzzy set theory) developed here nay be directly applicable or extended to more particular problems where some sort of decomposition nust be introduced to alleviate the difficulty in analysis, design, or evaluation. Clustering can also be made on the basis of some decision fu nct ions.