Modelling of human optimization

Use this forum to discuss any issues relating to Systems Dynamics and Systems Thinking.
Post Reply
patrick_at
Junior Member
Posts: 4
Joined: Wed Jan 11, 2012 9:47 am

Modelling of human optimization

Post by patrick_at »

Hi all,

System Dynamics models often incorporate humon decisions, based on information. For example a worker or a company that runs short on time for a project might increase working hours. It is clear that this works for a while, however it might decrease productivity in the long run. You could build such a model and analyze it, especially you could try to optimize.

However, what if you wanted to model the optimization procedure of the company itself? They might know about the loss of productivity. A simple procedure could be, that the company would increase working hours gradually, and then they after a while they could compare the output with the output before the increase of working hours. If output is falling, they could again decrease working hours, compare new output with old output, and accept the decrease in working hours if the new output is now better.

In an agent based model, agents could adapt their behavior exactly in this way, with discrete events where they compare the new value of their objective function with the old value. However, what would be the equivalent formulation in System Dynamics, if people in the system try to optimize their behaviour but do not really know which behaviour will lead to better results (actually they just try what works)?
Post Reply