A density function can be formulated with the maximum entropy method (MEM) that achieves entropy while meeting various constraints. The following three constraints are considered in the present work: a specified mean a specified variance, and knowledge that the density is one-sided (i.e., positive domain). A 1D gradient search (i.e., a line search) for obtaining the maximum entropy density is presented. Entropy and performance analysis results on a manufacturing system example are presented to compare the s-transition and maximum entropy density functions (MEDFs). The similarity between MEDF transitions and s-transitions is shown.
展开▼
