Compact evolutionary algorithms (cEAs) are optimization algorithms that require minimal computational cost. They do not require the storage of the population but they represent it by a distribution function. In all known cEAs, normal probability of density function (N-PDF) is used. In this paper, in order to improve the performance of cEAs and to reduce their complicity, we propose a more simple distribution to represent the population. The proposed distribution is the uniform probability of density function (U-PDF). Thus, our proposals are called uniform compact evolutionary algorithms (UcEAs). The aim is to investigate the influence of the distribution function on the performance of cEAs. Six new algorithms are proposed and denoted. In order to evaluate their performance, an important numerical experiment is performed. The algorithms are benchmarked on the IEEE CEC2014 functions. Furthermore, the nonparametric Wilcoxon's rank sum test and Freidman's test are used to compare and to rank the algorithms. The obtained results show that U-PDF is suitable to implement some cEAs. The time processing is commendably reduced by means of U-PDF.