AI could be the future of decision-making in the water industry

Posted 9 August 2017

Artificial intelligenceA South Australian researcher has developed a way to use artificial intelligence (AI) to solve stormwater system management and other problems facing the water industry.  

Dr Michael Di Matteo, a postgraduate student at the University of Adelaide, said his method could provide an optimised approach that yields more cost- and resource-efficient results than traditional approaches. 

Organisations traditionally use a multi-criteria analysis approach that allows humans to rank and score individual projects before choosing which ones to prioritise, Di Matteo said. 

But a new method utilising AI could provide an optimised approach that produces better results with less funding and fewer resources.

“We found that [using an AI optimisation method] you can get twice the stormwater harvesting capacity for half the cost, twice the water quality improvements for half the cost and twice the amenity score as judged by the local councils for half the cost.” 

These benefits were found after Di Matteo used AI to find an alternative solution to a funding allocation decision that needed to be agreed upon by three Victorian councils. 

“They wanted to fund 70% of their projects, and finding a combination of projects that gave them the greatest benefit for the lowest cost was a hassle,” said Di Matteo, whose research focuses on applying AI techniques to stormwater system problems. 
 
“There were millions and millions of possible combinations.” 

Di Matteo applied an ‘optimisation algorithm’ that trawled through the possible combinations and identified the ones that shared costs and benefits among councils in the most equitable and efficient way. 

The research was conducted after a decision had already been made by the councils involved, and Di Matteo is quick to stress that this is just a theoretical approach that is yet to be applied in practise. 

“It’s what you theoretically can achieve if you have the decision maker buy in to get there. It can be a useful benchmark when people are deciding which projects to go ahead with,” he said. 

Di Matteo will present on how to balance outcomes among various stakeholders using optimisation and visual analytics at the upcoming SA State Conference. To learn more, click here