保存的数据文件*.xls,是excel文件，里面是每步的数据,
要获得精确的数值，在algorithm中选择slow,
以下来之online:
To perform an optimization, you can use one of the algorithms below:
Gradient based algorithm
This algorithm should be used first to perform a local search. Based on the calculation of a local slope of the objective function this algorithm will use a parabolic approximation and jump to its minimum or use an iterated exponential step descent in the direction of the minimum.
If the properties of the objective function are known (continuous, differentiable at all point), then the gradient can be used straight on. It is usually faster than the Simulated Annealing algorithm.
Simulated Annealing based algorithm
This algorithm is a global stochastic search algorithm hence two successive runs of this method might not lead to the same result. It performs a global search that evolves towards local searches as the time goes on.
It is usually used to explore non-linear, multi-modal functions. These functions can also be discontinuous.
If the shape of the objective function is unknown, it is recommended to start with a Simulated Annealing then refine the results with a gradient descent. This approach is slow but works for a larger amount of functions.
A good way to quickly reach a solution when using the Simulated Annealing consists in specifying a low number of consecutive iterations without improvements (15 or 20). |
发表于 2002-12-26 23:09:33
发表于 2002-12-27 08:50:25
发表于 2002-12-27 08:52:07
