The inmate assignment project, in close collaboration with the Pennsylvania Department of Corrections (PADoC), took five years from start to successful implementation. In this project, we developed the Inmate Assignment Decision Support System (IADSS), where the primary goal is simultaneous and system-wide optimal assignment of inmates to correctional institutions (CIs). We develop a novel hier- archical, multi-objective Mixed Integer Linear Optimization (MILO) model, which accurately describes the inmate assignment problem (IAP). The IAP is the mathematical optimization formulation of the problem every correctional system faces which is to assign inmates to CIs and schedule their programs, while all legal restrictions and best practice constraints are considered. By using real inmate data sets from the PADoC, we also demonstrate that the MILO model can be solved efficiently. IADSS enables PADoC to significantly reduce the population management costs, and enhance public safety and security of the CIs. To the best of our knowledge, this is the first time that Operations Research (OR) methodologies have been built directly into the routine business practice of a correctional system, and used to optimize its operations. This successful project opens a rich and untouched area for the application of OR and optimization methodology. The new model and methodology can be utilized for the assignment of inmates in any correctional system.
Citation
ISE Technical Report 17T-013, Industrial and Systems Engineering Department, Lehigh University, 200 W Packer Ave, Bethlehem, PA, October 2017