This study models the compaction of soil by a vibratory roller compactor and examines changes to current designs that may provide more efficient compaction. The modeling of the soil differs from previous analyses of the compaction process in its use of a distributed-parameter characterization of the soil mass and in the application of nonlinear constitutive relations that predict the continuous evolution of residual stresses in the soil. The model was used to determine whether compactor performance might be improved by changes in the forward speed of the compactor or by redistribution of the weight of the frame. No improvement was found to occur. The model was also used to estimate the effects of varying the frequency of vibration to match the evolutionary changes in soil properties during compaction. Hardly any improvement over operation at constant frequency was indicated, but this finding may reflect the tendency of the model to underestimate the rate of stiffening in the soil. / Master of Science