摘要：AbstractDiscrete-event simulation of hospitals typically specifies flow by means of a process graph through which patients are routed. While this is generally fine for models in which processes are clearly defined, e.g. smaller units such as emergency departments, it falls short of capturing the fact that a patient can in principle go from one unit to any other unit if medical procedure requires it. This problem becomes even more evident when modeling a whole hospital, at which level specifying each individual treatment through directed edges and nodes becomes unviably complex. In the past three years, we have been developing a combined hospital simulation/space design tool in which processes are defined individually by patient, as sequence of visited units imported from the Hospital Information System (treatment chains). The major advantage of this approach lies in the fact that we can now design a future space layout in which assumed capacities (staffing, equipment, required space) can be tested against the recorded patient volume, using an Agent-Based Simulation to re-enact each treatment chain. In contrast to other hospital simulations, this approach targets early stages of architectural conception, during which the actual building structure is elaborated. Using the method, we can compare and contrast different layouts during that stage, making simulation a driver for design rather than a tool for late optimization within the final floor plan.
