This September researchers from the HOWS team showcased the Augmented Evolutionary Intelligence (AEI) system at the 17th International Computing & Control for the Water Industry Conference hosted by the University of Exeter.  Delegates got to try their hand at optimising water distribution networks using the new virtual reality based AEI system which employs cutting edge visualisation, machine learning and evolutionary computation to harness the power of human intuition to solve engineering problems.

The HOWS team presented two papers at the 17th International Computing & Control for the Water Industry Conference at the University of Exeter, UK.

Herman Mahmoud presented a paper entitled Generalising Human Heuristics in Augmented Evolutionary Water Distribution Network Design Optimisation. The paper proposes a methodology for capturing and generalizing engineering expertise from humans through machine learning techniques and integrating the resultant heuristics into an evolutionary algorithm to find optimal water distribution network designs.

The second paper entitled A Diameter Probability Distribution Genetic Algorithm for Least-cost Water Distribution Network Design was presented by Matt Johns. The paper describes a new search space reduction technique which uses data gained from previous evolutionary algorithm runs to improve the performance of an algorithm on larger, more complex water distribution network design problems.

This July the HOWS team presented two papers at the Genetic and Evolutionary Computation Conference (GECCO 2019) in Prague, Czech Republic.

Matt Johns presented a paper entitled Augmented Evolutionary Intelligence: Combining Human and Evolutionary Design for Water Distribution Network Optimisation in the Real-World Applications track. The paper describes recent work conducted on the HOWS project which combines user interaction, machine learning and evolutionary algorithms to improve the way water distribution network optimisation is conducted.

Nick Ross presented a paper entitled Human-Evolutionary Problem Solving through Gamification of a Bin-Packing Problem at the Interactive Methods workshop (iGECCO 2019). The paper explores the use of gamification as a mechanism to extract problem-solving behaviour from human subjects through interaction with a gamified version of the bin-packing problem, with heuristics extracted by machine learning. These heuristics are used to augment an evolutionary algorithm to improve search performance.