AbstractÐWhile gaining widespread acceptance among decision makers, simulation methodology still suers from its traditional de®ciencies. Simulation modeling lacks a well established science base. During the traditional model-development process various system assumptions are hardwired into the model code. The modi®cation of these assumptions usually requires major model overhauls with extensive programming eorts, resulting in minimal ease of model development and code reusability. This paper explains the development of a new uni®ed framework for simulation that includes an integrated methodology based on a formalism and science base deployed in an object-oriented software environment. For modeling ¯exibility, modularity and realism, it is necessary to separate logical and physical events, processes and activities of a system. In this new perspective, the modeler will perceive the system of interest in terms of several physical entities, augmented by some number of decision-making elements, each capable of making rational decisions that aect the behavior and contribute to the collective state of the system. The modularity of the user's conceptual model in terms of physical and decision-making entities with their programmed and non-programmed decisions and the capability of modeling and abstraction of these entities and processes at dierent hierarchical levels can vastly improve simulation's role as a decision-making tool.