Modern flour mills continuously mix up to 80 intermediate flour streams into 3–6 final streams, whereby each intermediate stream can be allocated to only one final stream. All streams are characterized by several intensive properties and a yield. The decision for mixing allocation involves maximizing the total sale value of all final streams, such that each of them corresponds, with respect to its property ranges, to a flour that can be sold on the trade market. The decision must thereby respect any constraints on the allocation variable and on the yields of the final streams. In this work, a tool is developed that automatically performs the discrete optimization for the allocation problem and outputs the best mixing decision for the given situation. Using heuristics-based optimization, the tool respects all constraints, selects the most lucrative combination from all possible traded flours, and determines the allocation of each intermediate stream to one of the selected traded flours. The optimization method can accommodate nonlinear mixing rules. The tool can trivially be extended to other mixing allocation problems, e.g., to select input streams with minimal total cost to mix to a single output stream with specified property and yield ranges.

This content is only available via PDF.
You do not currently have access to this content.