A Fast Method for Generalized Starting Temperature Determination in Two-Stage Simulated Annealing Systems

Simulated annealing is a stochastic process that has proven to be an effective method for approximating globally optimum solutions to many types of combinatorial optimization problems, especially in the field of VLSI computer-ended design. The major drawback to the simulated annealing algorithm is its typically very long running times. Several methods have been proposed for accelerating the simulated annealing algorithm. One method is to replace a significant portion of the stochastic operations with a fast heuristic. Simulated annealing can then begin from a lower starting temperature——-a latter stage of the algorithm - ~to further improve the solution produced by the heuristic. This paper presents a method for approximating this starting temperature in general, as well as experience with two-stage systems for solving the VLSI partitioning, traveling salesperson, and minimum—length rectilinear Steiner tree problems.
Note: Abstract extracted from PDF file via OCR

Varanelli, James, and James Cohoon. "A Fast Method for Generalized Starting Temperature Determination in Two-Stage Simulated Annealing Systems." University of Virginia Dept. of Computer Science Tech Report (1993).