Improving short-term water demand forecasting using evolutionary algorithms

Research question

Can a method for extracting weekly seasonality improve short-term water-demand forecasts in a district metered area?

Method and findings

Stańczyk and colleagues combined linear regression with evolutionary strategies and compared forecasts with support vector regression, multilayer perceptron and random forest methods. The case used a Wrocław district with multifamily housing. The authors report some daily forecasting cases with mean absolute percentage error below 2%. That figure is a result from their study, not a general accuracy promise.

Limitations

Results depend on the dataset, forecast horizon and local consumption patterns. The authors discuss the difficulty of comparing models across different operating conditions.

Engineering interpretation

Continue to the demand forecasting workflow.

Sources & further reading

• View original research: Improving short-term water demand forecasting using evolutionary algorithms ↗Stańczyk, Kajewska-Szkudlarek, Lipiński & Rychlikowski · 8 August 2022
  External source · Checked 24 September 2026

Source findings are distinguished from editorial interpretation. Apply current local criteria and project evidence when making engineering decisions.