Controlling resource requirements

A computer-implemented method of estimating resource requirements in an environment, the method comprising a preparation phase and a simulation phase, wherein the preparation phase comprises a machine learning training phase and a clustering phase, wherein the machine learning training phase comprises:

• receiving historical environment data from a plurality of distributed first databases, wherein the historical environment data comprises historical values of a time series for a plurality of features reflecting characteristics of the environment;
• receiving historical resource requirement data from at least one second database, wherein the historical resource requirement data comprises historical values relating to a resource requirement;
• and
• training a machine learning model to predict a resource requirement, wherein during the training phase a subset of features from the plurality of features is extracted;
• wherein the clustering phase comprises:
• determining at least two clusters in the historical environment data, wherein a cluster represents time periods having correlated values for features in the subset of features;
• determining, for a cluster, a correlation coefficient of the features in the subset of features;
• and
• determining, for a cluster, at least one identifying parameter of a distribution of the feature values of the subset of features;
• wherein the simulation phase for a cluster comprises:
• determining a distribution for a feature of the subset of features according to the correlation coefficient and the at least one identifying parameter for the cluster;
• selecting at least one value for the feature of the subset of features based on the determined distributions;
• and
• inputting the at least one value for the feature of the subset of features to the machine learning model to estimate a resource requirement in the environment in at least one time period the future.