A first experiment in logistics system simulation

This paper essentially describes the activities of the RAND Logistics Systems Laboratory during its first year of operation. During this year, the laboratory conducted an important first experiment in which it tested a series of proposed changes to the Air Force supply system. The test involved the creation of two versions of an Air Force logistics system, one which simulated a 1956 Air Force logistics system and another which was modified to contain the proposed supply policies. These two systems were operated in parallel and were given the same support task to accomplish. The system which was able to accomplish its job at lesser cost was deemed the better system. The simulation was a composite of man‐machine models, in which the men were assigned management responsibilities and functions and the machines performed routine actions. The men were identified by titles which have counterparts in the real‐world logistics system. The two logistics systems were operated under compressed time so that each hour in the laboratory simulated a day's experience in the real world. This enabled the simulation to cover 3–1/2 years' experience and 2 simulated wars. The experiment indicated that, according to the criterion established, the logistics system containing the newer supply policies was a better system. The experiment also indicated ways in which revisions to the procedures by which the new policies are implemented would benefit their future use. The experiment indicated that this type of logistics system simulation did develop many more insights into the operation of proposed policies than traditional analytical techniques normally provide, and it did extend the traditional descriptions provided to real‐world organizations of such proposed policies. At the same time, the experiment did highlight the complexities of testing logistics policies, even under the idealized conditions of the laboratory, and the problems of interpreting experimental findings.