Parallel Implementation of the Social Forces Model

We demonstrate how the use of a multicomputer can greatly accelerate the speed of a pedestrian movement simulator based on the social forces model. Our objective is to develop a simulator that updates the position of every pedestrian in real time; that is, 30 times a second. We have achieved this goal through the use of multiple processors. We describe the design of our parallel pedestrian movement model and present benchmark results demonstrating that 11 processors can update the positions of 10,000 pedestrians in about 1/50 of a second. The parallel algorithm is highly scalable, meaning that adding processors will enable the simulation of even larger crowds.