Dissipative channel dependence of monomial warm inflation: An effective Bayesian ranking
Phys. Rev. D 114, 103532 – Published 18 November, 2026
DOI: https://doi.org/10.1103/xm9t-6njs
Abstract
Within the warm inflation framework with the monomial potential , model branches are labeled by , where distinguishes the linear () and cubic () dissipative channels. The dependence of the effective Bayesian ranking of the potential branches on the dissipative channel is systematically examined. The linear and cubic channels are compared using the same background evolution, pivot scale prescription, compressed observable likelihood, and structure conditioned priors. The observables , , and are calculated with the channel matched scalar spectrum growth factor, while nested sampling is used to determine the posterior distributions and effective Bayesian evidence of the six candidate models. The mean log evidences from three independent random seeds give and for (4, cub) and (3, lin), respectively, relative to (4, lin). These three models form a closely spaced high evidence group. The corresponding values for (2, lin), (3, cub), and (2, cub) are , , and . Thus, the quartic potential remains competitive in both dissipative channels, whereas the support for the cubic potential depends strongly on the dissipative channel, and the quadratic potential is generally disfavored. Representative high evidence solutions lie within and . The lower evidence branches are constrained mainly by an overly large spectral index and by the limited prior volume of their high likelihood regions. The main evidence hierarchy remains stable under variations of the pivot e-fold number, prior ranges, independent reruns, and the width of the soft tensor penalty. The standard deviation of from run to run does not exceed 0.200, although the detailed ordering within the high evidence group is not statistically decisive. Reoptimization and evidence calculations for different thermal occupation fractions show that (4, lin) retains a good local fit without thermal occupation, whereas higher inflaton occupation fractions are strongly preferred by the three cubic branches. Within the adopted likelihood, priors, and spectrum prescription, the evidence ranking reflects both the compatibility of the spectral predictions with the observational constraints and the prior volume of the high likelihood region. The thermal occupation tests reveal different dependences on occupation for the two dissipative channels.