SAS Practice Prize

To do it successfully, three key things must be decided: what jobs should be done, who should do them and what sequence they should be done in – a variant of the "travelling salesman" problem in computer science. "The actual task of driving around and fixing roads can be quite time consuming," David says. "And if you don't get it right, you end up not being so productive. There's a lot of complications if you think about how many things could go wrong on a road – barriers or potholes, litter and all that stuff." A trained human scheduler takes two to three hours to manually set up a reasonable route, according to Elton. But thanks to the tool they developed, that time is cut down to 20 minutes – a sixfold reduction in planning time.

Although many off-the-shelf vehicle routing tools are available to tackle the problem, David and Elton say they can be costly, inflexible, include unnecessary features, or difficult to adapt for roadworks scheduling. "We have empowered our planning leads and our supervisors to spend time on more impactful work rather than spending hours every single day creating this plan," says Elton, a Strategic Improvements Analyst with Downer who also holds a degree in Mathematics and Statistics from the University of Melbourne. "It's a daily task that we've automated into a 20-minute tool that they can use right in the background – go get a coffee, come back and the job is much easier."

The productivity growth is passed on to road users in the form of safer roads, the pair say. After the tool was implemented, preliminary analysis showed a 25 percent increase in jobs completed, a 15 percent higher crew utility and 50 percent reduction in critical jobs. As an added benefit, optimising the road maintenance schedule also led to a reduction in carbon emissions. "We're quite a heavy transportation-based company," says Elton. "We do a lot of driving with a lot of trucks and a lot of fuel. So, one major outcome is that we now do it more efficiently, which means much less CO2 emissions."