Optimization under Uncertainty
Dealing with Risk is rather a psychological than a mathematical challenge. Researchers and practitioners, therefore, need to tackle personal and subjective issues, when dealing with decision making and optimization problems, and to step back from applying mathematical approaches only.
Of course, mathematical optimization approaches can help make analyses and improve decision making quality in uncertain situations, but successfully resolving uncertainty means to efficiently improve the quality of information. This, however, requires to understand why results are uncertain, and to find out how “nature” works with regard to the optimization problem. This is, in fact, smart reasoning and doing some sort of lab work, but it’s not a mathematical algorithm.
The process of uncertainty resolution aims at resolving uncertainty iteratively.
- Check if a safe decision is possible, and exclude inferior alternatives
- If a safe decision is not possible, resolve uncertainty about underlying problem, and adapt the underlying model to new knowledge
- Go back to the decision making stage, and start the process all over again, unless the optimum alternative is identified, or until you cannot truly say, which alternative is best
Check my publications to find a more detailed explanation of this process, and it’s application in real life decision making situations in chemical product/ process development projects.