Each affords different insights into agency performance. Queueing theory focuses upon single activities and can be used to develop reasonable standards for time to complete those activities and agency capacity. A disadvantage of queueing theory is that the mathematical model may not represent the behavior of the offender and the probation officer. Linear programming provides insight into what operational constraints cost society in terms of benefits or effectiveness sacrificed. As with queueing theory, it can be applied to a single agency but can also be used to make performance comparisons across agencies. Linear programming requires more data than queueing models. Some linear programming models can be solved by hand, but most are better handled by computer. Cost functions can be used to make efficiency comparisons and provide insight into which variables have the greatest impact upon cost. Cost functions require multiple observations of an agency over time, observations over many agencies, or observations over several agencies over time. The amount of data required and the complexity of the statistical modeling makes it the most costly of the three techniques to use. Fifteen footnotes are provided. (Author summary modified)