Industry News

Gurobi software drives award-winning inmate assignment project

A team from Lehigh University using Gurobi was awarded one of the top international prizes in operations research, the Daniel H. Wagner Prize, for its work developing the Inmate Assignment Decision Support System (IADSS). The award was announced at the 2017 INFORMS Annual Conference held in Houston. For more on the Wagner Prize, see page 47.

Assigning inmates to correctional facilities is a complex process taking into account literally dozens of factors relating to the inmate, such as criminal history, demographic characteristics, and mental and physical health needs, as well relating to the prison system, such as facility utilization levels, support program availability and existing inmate characteristics at each facility. Good assignments are important for both the inmate, since they can result in a lower chance of violent interactions with other inmates, as well as faster access to treatment programs, and the prison system since it can reduce staff workload and the number of prison transfers.

The application the Lehigh team developed took five years to create but saved the Pennsylvania Prison System $3 million in just the first year of use. It has reduced the number of incidences of inmate violence, reduced transfer rates and staff workload, and has also reduced the time it takes for inmates to get access to treatment programs. What literally took a staff of seven a full week to do is now done in just a few minutes on a daily basis by the application. The team’s process and results are detailed in the paper submitted for their winning entry, “The Inmate Assignment and Scheduling Problem and its Application in the PA Department of Corrections.”

The problem is modeled as a mixed integer linear optimization problem where the objective is a hierarchically weighted sum of six different objectives including penalizing violations of assignment criteria, not exceeding capacity constraints and minimizing wait times for treatment programs. The model contains about 30,000 binary and about 200 integer variables, and has about 25,000 constraints. Gurobi is used as the engine with the solve terminated when it reaches a small predetermined optimality gap. In just a few minutes, Gurobi is able to provide a high-quality solution, and the application is now in daily use at the Pennsylvania Department of Corrections.

The project participants included Kristofer B. Bucklen from the Pennsylvania Department of Corrections and professors Tamás Terlaky, Louis J. Plebani and George R. Wilson and grad students Mohammad Shahab-safa, Naga Venkata Chaitanya Gudapati and Anshul Sharma from Lehigh’s Industrial and Systems Engineering Department.

AIMMS picks UniSoma as its AIMMS Partner of the Year

AIMMS named UniSoma as its AIMMS Partner of the Year for 2017. Founded in 1984, UniSoma has played a leading role in the introduction of operations research and analytics at some of Brazil’s largest companies, including JBS, Suzano, Aurora Alimentos, Marfrig, Itambé and many others. The company develops advanced planning and complex scheduling solutions, helping businesses save millions with supply chain optimization. AIMMS has accompanied UniSoma throughout this journey. The companies formalized their partnership in 2009 and have been working together for 20 years.

UniSoma received the award for demonstrating business excellence and customer satisfaction while delivering transformative AIMMS solutions. AIMMS also recognized UniSoma for excellence in quality, value, services and innovation. The award highlights the strength of the partnership and UniSoma’s focus on accelerating optimization projects for clients.

Gertjan de Lange, AIMMS senior vice president, Connecting Business & Optimization, said: “This award recognizes the business value and innovation that the partnership is delivering to our joint customers. We have created and nurtured this value with UniSoma’s deep business and industry expertise and AIMMS’ market-leading prescriptive analytics technology. It’s a true pleasure to honor UniSoma’s achievements and their contribution to the expansion of AIMMS’ footprint with this award.”

AIMMS created the Partner of the Year Award in 2015 to recognize the valuable contribution of its extensive network of partners. The selection process is based on the following criteria: generating business opportunities, delivering customer success, leveraging AIMMS technology and boosting AIMMS’ Footprint.

Luciano de Moura, commercial director at UniSoma, said: “The award is the recognition of two decades of partnership between UniSoma and AIMMS. During this time, we jointly helped leading companies in Brazil to improve their planning decisions and to achieve new levels of efficiency. UniSoma and AIMMS share values such as focus on customer results, innovative attitude and commitment to the quality of our SCP Solutions. We will continue to work together to increasingly deliver state-of-the-art analytical solutions to the market.”

What’s new in Analytica 5.0?

The Lumina team recently announced the release of Analytica 5.0, a major new release that includes literally hundreds of new and improved features. Users with active support can upgrade free to a subscription to Analytica 5.0. Others can purchase a new subscription.

Highlights include:

  • New toolbar icons and matching pop-up node controls on diagrams give faster access for common operations on each node.
  • Multithread computing: Parallel computing speeds up the most common calculations using multithreading when your computer has multiple processors or cores.
  • Cell table formats: Users can now set formats in tables, including text style, size, color and font, borders, background colors, number format and images. Users can specify formats by cell, headers, rows, columns, slices or ranges, and they pivot along with the table view. Conditional formats let users compute styles or colors to create heatmaps and other powerful visualizations.
  • New Find dialog: Click or press control+f for a vastly more powerful Find dialog that lets users search their model by object identifier, title, description and other text attributes. It can also search the online Analytica wiki documents.