2013/07/25 by E. Canessa, Canessa, E.
Physics and Astronomy · #Data Analysis #FOS: Physical sciences #Statistical Mechanics (cond-mat.stat-mech) #Statistics and Probability (physics.data-an) #cond-mat.stat-mech #physics.data-an
paper · pdf · doi:10.48550/arxiv.1307.6681
arxiv created 2014/01/20 · arxiv updated 2014/01/21
The log-periodic equation for the entropy S = - (k/a) ∑i=1N pi sin(a ln pi), based on the forgotten Sharma-Taneja entropy measure, is studied for the first time with N the total number of system states and pi the associated probabilities. It is argued that an oscillatory regime for S could in principle be understood in terms of a linear time-dependent behavior for the associated probabilities in analogy with a spring system of frequency w gaining momentum from the surroundings. The physical meaning for the production of entropy given by the parameter a relates the angle ωt. We discuss its properties, and make a connection with the non-extensive Tsallis, Rényi, Boltzmann-Gibbs and Shannon entropies as special limiting cases for systems with constant mass. Our non-trivial form of entropy displays peculiar concavity and addition properties. Log-periodicity in S, concavity lost, and non-additivity are manifested by increasing the value of the coefficient a, which sets the variations with respect to the behavior of the monotonic Gibbs entropy function. We have considered systems with linearly increasing mass and have explained how "oscillating entropy" in such systems may appear.