An Optimization-based Algorithm Can Reduce Deadhead Distance by 13.2 Percent for Winter Road Maintenance Operations in Northwest Iowa.
Researchers Apply Optimization Algorithms to Design Optimal Routes for Winter Road Maintenance Operations Using Data from Iowa DOT District 3.
Made Public Date
03/30/2021

896

Iowa
United States
Identifier
2021-01546
TwitterLinkedInFacebook

Iowa DOT Office of Maintenance Snowplow Optimization

Summary Information

Winter road maintenance activities include removal of snow and ice from roadways and spreading materials to provide safe and efficient service for road users. Iowa DOT District 3 is responsible for servicing about 4,000 lane miles from 20 depots for winter road maintenance. Two types of maintenance trucks with different capacities (12,000 or 16,000 pounds of solid material) are primarily used in Iowa. The current routes for District 3 are designed based on staff knowledge and past experience. To minimize the deadhead distance (travel while not actively treating roadway) and meet service expectations, two optimization problems were identified and solved. The first problem was to design the optimal truck routes for winter maintenance for single depots under the district’s current responsibility maps. The second problem was to design routes for multiple depots with intermediate facilities. The study used data from two winter seasons, from October 1, 2016, to April 1, 2017, and from October 1, 2017, to April 1, 2018.

Methodology

Weather data was collected for this study, specifically daily snowfall and daily snow depth. In addition to weather, operational data of each maintenance truck was collected, including the GPS location, plow position, spreading rate and type, truck speed and direction, truck ID, and timestamp.

The first problem is formulated as a single-depot winter maintenance routing problem. The second problem is formulated as a multiple-depot winter maintenance routing problem, where trucks could reload material at locations other than their home depot. Both optimization problems were solved computationally, considering the constraints of road segment service cycle time, vehicle capacities, fleet size, road-vehicle dependency, and work duration. A sensitivity analysis for the spreading rate, the number of pounds of material per lane mile, is assessed to check if it changes the optimized routes.

Findings

  • The results from solving the single-depot optimization problem show a 13.2 percent reduction in deadhead distance compared to current operations. The reduction in deadhead distance could be even larger because while the optimized routes strictly satisfy all constraints, the current operations might not.
  • The total optimized travel distance for all sectors under the multiple-depot scenario is 4,859.8 miles, while the single-depot scenario is 4,919 miles. The deadhead distance savings in the multiple-depot scenario compared to the single-depot scenario differ by 1.2 percent, which was determined to be insignificant.
  • The sensitivity analysis of the spreading rate parameter shows that this parameter only impacts routes that service roadways with a service level of C, which represents low travel demand roads.

 

Goal Areas
Results Type
Deployment Locations