An Analysis of a Randomized Local Search Algorithm for the Entropy Space

Abstract

In previous work, we proposed a randomized local search algorithm to determine the distribution associated with a vector given in the entropy space. The algorithm also finds the nearest vector and corresponding distribution if the given vector is non-entropic. The utility of the algorithm for entropy optimization has been shown. A trade-off between entropy functions is observed. The convergence is better in comparison to the other algorithms used in the entropy region. Finally, the time and space complexities have been computed.