The most flexible DOE approach is called optimal design of experiments. That approach allows you to create tailor-made experimental designs, taking into account any budget constraints and logistic difficulties. Except in those rare case where an optimal experimental design is known with certainty, tailor-made experimental designs are computed using heuristic optimization algorithms such as coordinate-exchange algorithms, point-exchange algorithms or metaheuristics methods such as simulated annealing.
While these algorithms generally produce good results, they cannot guarantee that they find the very best design for your case. As a result, the design that you get from your favorite software may be worse than the one your competitor gets for the same problem. The design you get today for a given scenario may be worse than the one you got yesterday.
๐ข๐๐ฟ ๐๐ผ๐ณ๐๐๐ฎ๐ฟ๐ฒ ๐ด๐๐ฎ๐ฟ๐ฎ๐ป๐๐ฒ๐ฒ๐ ๐๐ต๐ฎ๐ ๐๐ต๐ฒ ๐ฑ๐ฒ๐๐ถ๐ด๐ป ๐๐ผ๐ ๐ฐ๐ฟ๐ฒ๐ฎ๐๐ฒ ๐๐ผ๐ฑ๐ฎ๐ ๐ถ๐ ๐ฎ๐ ๐น๐ฒ๐ฎ๐๐ ๐ฎ๐ ๐ด๐ผ๐ผ๐ฑ ๐ฎ๐ ๐ฎ๐ป๐ ๐ฑ๐ฒ๐๐ถ๐ด๐ป ๐ฐ๐ฟ๐ฒ๐ฎ๐๐ฒ๐ฑ ๐ฝ๐ฟ๐ฒ๐๐ถ๐ผ๐๐๐น๐ ๐ณ๐ผ๐ฟ ๐๐ผ๐๐ฟ ๐ฝ๐ฟ๐ผ๐ฏ๐น๐ฒ๐บ.This is possible due the cloud-based nature of Effex.
Effex keeps track of all good designs generated by its users, and it stores the very best ones. If you want to create a new design in Effex, then Effex executes its coordinate-exchange algorithm (as do other software packages) and compares the resulting design to the best design previously generated. In the final step, Effex returns the best of the two designs. This way, the DOEs on offer will continuously improve. Smart, isnโt it?
At Effex, it is our conviction that, when generating experimental designs for its user, we have to take advantage of 21st century technology. Rather than starting from scratch whenever a new design is needed, Effex exploits the advantages of cloud-based software to build on existing DOE knowledge and lets its users benefit from it!
