Heuristic Acceleration of Force-Directed Placement

Two heuristic methods are presented for accelerating convergence of a force-directed placement problem. The first stabilizes the derivative of the repulsion force. The second uses information on device movement and instability characteristics to make a predictive extrapolation. Convergence is accelerated by replacing iterations with the faster heuristic iterations. A standard implementation from the literature is made three to four times faster by using these techniques.