Extending a cognitive architecture with a representation of physiology allows one to simulate the effects of homeostasis on cognitive processes. The underlying physiological substrate in the ACT-RΦ architecture allows one to model potential interactions between homeostatic affect and cognition. In this paper, we de-scribe an extension of the ACT-R utility mechanism that uses information from the physiological representation provided by ACT-RΦ. We also discuss a model that uses this new system to simulate the effects of homeostatic-based thirst on rule utility. The model completes a variation of the ultimatum game that used a primary reward (water) as opposed to a monetary reward. In the experimental version of this game, researchers induced thirst in some participants by administering hypertonic saline via an intravenous line. Our model was able to represent both the physiological (i.e., hypertonic saline) and behavioral (i.e., forced selection of a primary reward) portions of the experiment. The model also can potentially provide predictions of behavior during the task that could not be observed due to the nature of the experimental condition.