Simulated Annealing Beats Metropolis in Combinatorial Optimization
نویسنده
چکیده
The Metropolis algorithm is simulated annealing with a fixed temperature. Surprisingly enough, many problems cannot be solved more efficiently by simulated annealing than by the Metropolis algorithm with the best temperature. The problem of finding a natural example (artificial examples are known) where simulated annealing outperforms the Metropolis algorithm for all temperatures has been discussed by Jerrum and Sinclair (1996) as “an outstanding open problem”. This problem is solved here. The examples are simple instances of the well-known minimum spanning tree problem. Moreover, it is investigated which instances of the minimum spanning tree problem can be solved efficiently by simulated annealing. This is motivated by the aim to develop further methods to analyze the simulated annealing process.
منابع مشابه
A Simulated Annealing Algorithm for the Satisfiability Problem Using Dynamic Markov Chains with Linear Regression Equilibrium
Since the appearance of Simulated Annealing algorithm it has shown to be an efficient method to solve combinatorial optimization problems such as Boolean Satisfiability problem. New algorithms based on two cycles: one external for temperatures and other internal, named Metropolis, have emerged. These algorithms usually use the sameMarkov chain length in the Metropolis cycle for each temperature...
متن کاملA Comparative Study of Genetic Search
We present a comparative study of genetic algorithms and their search properties when treated as a combinatorial optimization technique. This is done in the context of the NP-hard problem MAX-SAT, the comparison being relative to the Metropolis process, and by extension, simulated annealing. Our contribution is twofold. First, we show that for large and diicult MAX-SAT instances, the contributi...
متن کاملA Fast Algorithm for Simulated Annealing
We present a new deterministic algorithm for simulated annealing and demonstrate its applicability with several classical examples: the ground state energies of the 2d and 3d short range king spin glasses, the traveling salesman problem, and pattern recognition in computer vision. Our algorithm is based on a microcanonical Monte Carlo method and is shown to be a powerful tool for the analysis o...
متن کاملUsing simulated annealing to design good codes
A NNEAL, as defined in Webster’s New Collegiate Dictionary, means: 1) to heat (as glass) in order to fix laid-on colors, and 2) to heat and then cool (as steel or glass) usually for softening and making less brittle. The cooling stage of annealing is usually done slowly to reduce the number of defects in the crystal structure (as steel) and to minimize the potential energy stored in the molecul...
متن کاملSearching Randomly for Maximum Matchings
Many real-world optimization problems in, e. g., engineering or biology have the property that not much is known about the function to be optimized. This excludes the application of problem-specific algorithms. Simple randomized search heuristics are then used with surprisingly good results. In order to understand the working principles behind such heuristics, they are analyzed on combinatorial...
متن کامل