Industrial production: reducing waste

This exemplar is about reducing waste in industrial production. Reducing waste has more positive impacts as it will also reduces energy, puts less pressure on the environment, and helps reduce cost.

When does the waste appear in the production and why? In this exemplar, waste comes when moving from production process of one product to that of another. Whatever is produced during the  he transition period is not a product and is therefore waste. Thus, there is a considerable gain to be achieved if the transition period is minimised.

One way to educe the transition period is by improving the prediction of relevant measurements (to know when exactly a new product’s production is ready to start). This is what the present data set is about.

This data set is obtained from a real production, though it has been normalised and simplified, it still reflects a real situations.

The data set is a data matrix:

• First row has column titles A, B, …, H, X.
• Each subsequent row represents measurements from one point in time, but the rows come in random order so there is no information from consecutive rows.

The task is simply to predict the value in column X.

The data set has more than 11000 rows and can be handled any way the researcher finds appropriate according to his/her prediction method.

At the workshop we will present another data set for on-site calculation. The best prediction method will be the one that can make the closest predictions for that test set of ca 3000 data rows. Closeness will be determined by the root-mean-square deviation, , from the actual measurements.

The dataset is available here: industrialproductiondataset.csv .