Abstract: | This paper describes the performance of learning, markup and pseudo-maximizing pricing rules when applied using data on past operations in simulated environments. Performance is evaluated on (1) average last profit achieved after ten periods of experimentation with alternative prices and (2) total profit achieved over ten periods. The pseudo-maximizing rule performs at least as well as other feasible rules when the demand and cost curves are both deterministic and when the demand curve is deterministic and the cost curve is linear, but stochastic. In all other settings, the markup rule performs at least as well as other feasible rules. |