Optimizing inmate treatment program scheduling

Ramin Fakhimi, Mohammad Shahabsafsa, Anshul Sharma, Tamás Terlaky, and Luis F. Zuluaga

Jun 5, 2024

Pennsylvania currently houses approximately 40,000 inmates across 24 state correctional facilities. Prior to being granted parole, all inmates must complete a specific set of treatment programs. Certain programs can progress only at the pace of an individual undergoing treatment. That is, until one inmate completes the program, another must wait for their turn to begin. Other programs progress via a cohort, with a set number of inmates starting and ending the programs on the same days.

At issue is the fact that inmates can remain in the prison system longer than their minimum sentence if they miss treatment or are otherwise unable to access a required program. This can delay their return to society and result in increased demand for Department of Corrections resources and state funding. It is also extremely difficult and time consuming for corrections department officials to obtain a feasible and effective scheduling of the treatment programs for inmates.

A similar inmate assignment problem was previously solved by a group of graduate students working with Tamás Terlaky, professor of industrial and systems engineering at Lehigh University. Terlaky’s group was tasked with assigning up to a thousand inmates per week to the most appropriate prison facility. Graduate students Mohammad Shahabsafsa and Anshul Sharma incorporated the solution into their thesis work.

The deployment of their proposed solution method significantly reduced the time needed to complete this facility assignment task, resulting in millions of dollars saved for the state, improved safety, and a decrease in the prison population. Following the completion of this project, Terlaky, Shahabsafa, and Sharma founded the operations research and analytics consulting company Optamo LLC (Bethlehem, Pennsylvania).

Supported by PITA, Luis F. Zuluaga, professor of industrial and systems engineering at Lehigh University, and graduate student Ramin Fakhimi partnered with Optamo to investigate inmate treatment programs. The research team focused on tracking individual and group programs to optimize the treatment schedule required for parole. The team’s primary goal aimed to expedite the start of inmates' necessary programs and ensure they finish them within the shortest amount of time. A secondary goal aimed to reduce the amount of time program instructors spend holding classes, considering the number of inmates already in the program’s pipeline and the resources available.

Using data and scheduling procedures provided by the Pennsylvania Department of Corrections Office of Research and Development, the team developed a proof-of-concept mathematical optimization model for detailed scheduling of treatment programs for correctional systems using advanced prescriptive analytics and operations research methodologies. Led by Zuluaga, the team demonstrated that the best way to schedule the inmates and associated programs can be efficiently determined using appropriate data and mathematical optimization methods. Considering the large number of decision variables involved, as well as the number of inmates, the team believes their model will yield substantial benefits to a correctional system. A fully implemented system can assist in staffing support, monetary savings, security improvements, decreased prison population, reduced prison sentences, and an increased recidivism rate.

Leveraging the same methodologies, Optamo also developed solutions to optimize the assignment of inmates to facilities and housing units as well as to optimize inmate transportation. These program scheduling systems, as well as the other solutions, can be implemented at any state department of corrections in the United States, benefiting corrections departments and inmates nationwide, and helping Optamo grow into a thriving Pennsylvania business.