D. Lindström has developed an algorithm for optimization of expensive functions which is based on response surface methodology which means that a model of the expensive function is created. To optimize the response surface the algorithm is making use of a grid zoom. This grid zoom is replaced by a genetic algorithm which is based on biological evolution. Measurements are done before and after the modification focusing on CPU time. A vectorization of the code is also performed.

Expensive functions often appear in the real world. An expensive function could be to run a simulation with certain settings to see how good they are. Because the simulation takes a long time to run you can't just try everything but instead need to strategically select which parameters to test. This it what Lindströms algorithm does.

The results show that the grid zoom was more effective than expected but that it is ineffective in higher dimensions.