A function f (x), assuming that x=(a, b), a and b can be any floating-point, o a, b is equal to f (x) is the largest, with the search for a, b value steps to be larger, when found that f (x) larger search steps to reduce, know, search the maximum x also can be [a, b, c, d... And this algorithm is how to write?

CodePudding user response:

The original poster can study the simulated annealing algorithm, this algorithm can find the function in a certain range of the approximate optimal solution, there are a lot of application in engineering,

Generally greedy algorithm can find the local optimal solution, the simulated annealing algorithm under certain parameters, can jump out of local, find interval optimal solutions,

Baidu once its pseudo code or python implementation, see if you want,

CodePudding user response:

The original poster is the requirement that the optimal vector super parameters?