Bioreactor

This is one of a set of benchmark problems presented in "Appendix A: Challenging Control Problems" of Neural Networks for Control, Miller, Sutton & Werbos, Eds., MIT Press, 1990.

"The bioreactor is a tank containing water, nutrients, and biological cells, shown in Fig. 1. Nutrients and cells are introduced into the tank where the cells mix with the nutrients. The state of this process is characterized by the number of cells and the amount of nutrients. The volume in the tank is maintained at a constant level by removing tank contents at a rate equal to the incoming rate. This rate is called the flow rate and is the variable by which the bioreactor is controlled. The bioreactor control problem is to maintain the amount of cells at a desired level."

Fig. 1: The Bioreactor tank (from Neural Networks for Control, Miller, Sutton & Werbos, Eds., MIT Press, 1990, page 477)

At NWCIL, we used the DHP Adaptive Critic method to design a controller for this (and other) of the benchmark problems.

The Bioreactor animation was developed to assist the researcher to visualize the nutrient and cell quantities that result from application of the current controller being designed, following a disturbance in the nutrient and cell quantity ratio.

Each color shade in the animation represents a ratio of nutrient and cells.

• 'target' represents the desired set-point color shade
• 'nutrient' represents amount of nutrients in the tank
• 'cell' represents amount of cells in the tank
• 'control' represents the current inflow rate

The following animation demonstrates the action of one of the final controller designs during the training process:

Play the Bioreactor animation (flash) by Sangsuree Vasupongayya