Optimization methods evolutionary strategies

Selecting the right variables often improves the models and makes interpretation easier. When there are too many descriptors, and especially when these descriptors do not have a clear physico-chemical meaning (e.g., connectivity indices and other 2D descriptors), stochastic methods such as genetic algorithms and evolutionary strategies can be used for finding an optimal subset of descriptors [91,92]. 

Evolutionary Methods. The objective of all evolutionary methods for selecting and sequencing separation processes is a systemelic approach to an optimal system. These methods all mek to improve the existing flowsheet. Stephanopoulos describes the elements of an evolutionary strategy as follows ... 

Algoridunie Methods. The use of algorithmic approaches to process synthesis is, in theory, the only certain way to develop an optimal process. This is true because every possible process sequence can be considered rigorously. Unfortunately, because of the nature of the combinatorial problem that arises, even for relatively simple situations, the use of strictly algorithmic syiproaches is often unrealistic. However, when they are us in conjunction with heuristic and evolutionary strategies, they represent the best hope for rational process synthesis. 

Intuitively, one might imagine that the heuristics presented earlier are not of equal merit for all simations. Indeed, many of the process synthesis methods used today rely on the ordering of these heuristics either to generate an initial sequence that can be improved through an evolutionary strategy or to arrive at a final optimal sequence. 

---