Dispatch Optimization in Bulk Tanker Transport Operations

Friday, November 10, 2017

Photo - The Wagner Prize presentation team: (left to right) Ted Gifford, Daniel Vanden Brink, Tracy Opicka, Robert Randall and Ashesh Sinha.

I was proud to be part of a recent project that delivered significant, quantifiable economic savings as well as improvements in productivity and service. The project duration and Time to Payback (TTP) were rapid: less than four months elapsed from kick-off to initial solution deployment. In recognition of the operations research work and underlying mathematics, Princeton Consultants joined our client, Schneider, www.schneider.com, a leading transportation company, as finalists for the 2017 INFORMS Daniel H. Wagner Prize.

In transportation, it is a classic problem to find globally optimal or near-optimal solutions for the assignment of many assets—such as trucks, trailers and drivers—to freight movements. In Truckload (which features random one-way networks), the driver is the primary constrained resource and the trailer is generally treated as a commodity kept as a count in a pool (“a 53’ van is a 53’ van”).

Our project concerned the Bulk Transport (fuels/chemical) division, with a fleet of 1,000 tractors and drivers and 1,600 trailers, and pickup and delivery locations across the U.S. It transports more than 10,000 distinct commodities in a year. The division must treat its assets differently than does the Truckload division, which adds complexity to the assignment problem. The interaction properties of the commodities impose product-sequencing constraints and inter-order tanker wash and preparation processes, as well as selection of specific trailer configurations. In almost all cases, tanker trailers are routed to independent facilities where they are washed and prepared for their next use.

In the optimization modeler’s perspective, the drivers are still constrained, but the trailers are not just a commodity. A customer may reject a trailer if the previously hauled commodity, or if the last wash performed on the trailer, was objectionable.

Before our project, the prevention and management of trailer rejections was done manually. Agents had to read rules from customers in many different forms (fax, email, PDF, etc.) and in many cases with different terminologies (for example, one customer would refer to a chemical’s brand name, whereas another would use its generic name).

Our team developed a software tool for agents and customers to input the valid and invalid prior commodities for a given commodity. The inputs could be the manufacturer name, product family, or properties. Schneider valued the speed of this validation, and the rules engine we built was two orders of magnitude faster than required.

For the optimization model, we helped design a multi-phase, multi-dimensional matching algorithm to solve for the large combinatorics of the tank washes. We worked with the client team to develop new business processes that enable business planners to leverage optimized solution recommendations.

Schneider has calculated the benefit at $4 million in annualized operational and capital cost savings, as well as in significantly improved productivity and customer service that this new system has been delivering since implementation.

The submission process for the INFORMS Wagner Prize was rigorous and worthwhile. Schneider’s Dr. Ted Gifford, who has won the prize before, led the effort, managing the contributions from the team members. The judges provided initial feedback, asking for clarification or amplification about the mathematics and operations research techniques, the system deployment, and business objectives and benefits. Dr. Gifford took these all into account and led our presentation at the INFORMS Annual Meeting. The formal submission will be published in 2018 in Interfaces.