Optimal Repositioning of Empty Containers

Pacer Stacktrain

3-5%

Reduction in movement costs

Overview

Empty container movements result from geographical imbalances in supply and demand; for example, there are more loaded containers going from west to east in North America than from east to west, due to a high level of exports from China. We developed a model to help Pacer Stacktrain minimize empty container movements.
train8-case-study-img

Business Problem

Pacer Stacktrain was one of the largest intermodal companies in the United States engaged in the movement of containers. Companies like Pacer perform both loaded container movements (which generate revenue) and empty container movements (which do not generate revenue). An empty movement is used to reposition the container between two loaded movements. Pacer needed to decide how loaded container needs could be met by performing a minimum number of empty movements. It hired Optym to build a nationwide network planning model to determine optimal empty container moves.

Our Approach

We worked closely with the Pacer team to understand and document the business requirements of its network planning system. We modeled the movement of containers in a time-space network so every move in the real world corresponded to a move in this network.

We developed a mixed integer programming (MIP) model to determine optimal empty container movements. Then, we packaged the optimization model in a graphical user interface so network planners could easily use it. In the end, we validated model results by comparing them with those in the historical data.

Key Benefits

  1. The system generated a 3 to 5 percent reduction in empty container movement costs.
  2. By showing that optimal repositioning of containers could increase asset utilization, the system demonstrated that more loaded shipments could be moved using the same number of containers.

Results

Pacer Stacktrain used the model to perform what-if studies related to empty container movements and container fleet sizing and mix and to assess the impact of fleet size on service quality.