Abstract: The distribution estimation algorithms use statistical learning to create a probability model from a macro point of view, use the probability model to describe the distribution of problem solution in the solution space, and obtain the advantage of individuals by evolutionary computation. At present, the discrete distribution estimation algorithms are already quite mature, but receach progresses of continuous distribution estimation algorithm are slow. This article used the idea of uniform distribution to narrow the sampling field for continuous optimization problems, designed a new distribution estimation algorithm. Experimental data show that this kind of distribution estimation algorithms are valid.

Key words: Distribution estimation algorithms,Uniform distribution,Function optimization