G. Rudolph. Massively Parallel Simulated Annealing and its Relation to Evolutionary Algorithms. Evolutionary Computation, pages 361--382, 1994.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....in the following manner: First, one randomly chosen parent is reproduced. Second, the reproduced individual is mutated by adding a realization of a normally distributed random number with variance one and expectation zero. Hence, the resulting search distribution (the joint generating distribution [22]) is a multimodal mixture of Gaussians. the fitness values, it just allows the search strategy to adapt. V. Conclusion Neutrality is a necessity for self adaptation. Actually, the design of neutral encodings to improve the e#ciency of evolutionary algorithms is a well established approach: ....

G. Rudolph. Massively parallel simulated annealing and its relation to evolutionary algorithms. Evolutionary Computation, 1(4):361--382, 1994.

....is generated in the following manner: First, one parent is reproduced. Second, the reproduced individual is mutated by adding a realization of a normally distributed random number with variance one and expectation zero. Hence, the resulting search distribution (the joint generating distribution [22]) is a multimodal mixture of Gaussians. redundancy does not alter the distribution of the fitness values, it just allows the search strategy to adapt. V. Conclusion Neutrality is a necessity for self adaptation. Actually, the design of neutral encodings to improve the e#ciency of evolutionary ....

G. Rudolph. Massively parallel simulated annealing and its relation to evolutionary algorithms. Evolutionary Computation, 1(4):361--382, 1994.

....strategies (ES) Rechenberg, 1971# Schwefel, 1975) and evolutionary programming (EP) Fogel, 1966) are subclasses of evolutionary algorithms. Simulated annealing (SA) and single trial versions of ES are closely related in case they are designed for optimization over continuous variables (Rudolph, 1994). Following the definition of Back (1996) the general evolutionary algorithm can be defined as an 8 tuple EA= I# Phi# Omega # Psi#s# ##) # (20) where I = A x Theta A s is the space of individuals , and A x ,A s denote arbitrary sets. Phi = I ; IR denotes a fitness function (quality ....

Rudolph, G. (1994). Massively Parallel Simulated Annealing and its Relation to Evolutionary Algorithms, In: Evolutionary Computation, Vol. 1, No. 4, pp. 361-382.

....and evolutionary computation. Davis and Principe emphasized the relationship existing between the mutation probability and a temperature for the simple genetic algorithm [5] Suzuki [26, 27] analyzed Markov chain models of this algorithm using the analogy with simulated annealing. Rudolph [21], and Mahfoud and Goldberg [18] compared massively parallel simulated annealing and genetic algorithms. Van Nimwegen et al. 19] studied meta stability in the royal road genetic algorithm deeply with low mutation probabilities. The connection has been investigated by Cerf in a series of papers [3, ....

G. Rudolph. Massively parallel simulated annealing and its relation to evolutionary algorithms, Evolutionary Computation, 1, 4, (1993), 361-383.

....performance of SA and EAs in these results is mixed, which suggests that SA and EAs may be relatively superior on different classes of problems. Theoretical investigations into SA and EAs have focused on the development of hybrid algorithms that employ techniques developed for both SA and EAs [7, 20]. These authors argue that the strengths of these two classes of algorithms can be combined. However, their arguments do not provide a basis for comparing the performance of SA and EAs themselves. In this paper we theoretically analyze the relative performance of SA and EAs by comparing their ....

G. Rudolph, Massively parallel simulated annealing and its relation to evolutionary algorithms, Evolutionary Computation, 1 (1993), pp. 361--383.

....to survive for the next generation, while the other one is discarded. This simple selection mechanism is typically characterized by the term (1 1) selection. Obviously, this simple strategy does not work with a population of individuals, but rather has some similarity to simulated annealing [45]. A first population based, multimembered evolution strategy or ( 1) strategy was designed by Rechenberg to introduce a recombination operator, allowing the 1 parent individuals to form one offspring by recombination and mutation. The offspring eventually replaces the worst parent ....

G. Rudolph. Massively parallel simulated annealing and its relation to evolutionary algorithms. Evolutionary Computation, 1(4):361--382, 1993.

....which attempts to achieve Boltzmann distribution over the population. Mahfoud and Goldberg (1992) introduced a parallel genetic version of simulated annealing called parallel recombinative simulated annealing. This algorithm attempted to harness the strengths of both SAs and GAs. Recently Rudolph (1994) developed a Markov chain formulation of SAs and GAs for analyzing their similarities and differences. Jones and Stuckman (1992) made an interesting effort to relate GAs with Bayesian approaches to global optimization. They noted the similarities and differences between these two approaches and ....

Rudolph, G. (1994). Massively parallel simulated annealing and its relation to evolutionary algorithms.

....which attempts to achieve Boltzmann distribution over the population. Mahfoud and Goldberg (1992) introduced a parallel genetic version of simulated annealing called parallel recombinative simulated annealing. This algorithm attempted to harness the strengths of both SAs and GAs. Recently Rudolph (1994) developed a Markov chain formulation of SAs and GAs for analyzing their similarities and differences. Jones and Stuckman (1992) made an interesting effort to relate GAs with Bayesian approaches to global optimization. They noted the similarities and differences between these two approaches and ....

Rudolph, G. (1994). Massively parallel simulated annealing and its relation to evolutionary algorithms.

.... that GAs with a spatial population structure are not only easily to map onto massively parallel computers but also offer a better solution quality than traditional GAs [11, 5, 17, 12, 16] Recently, this approach was also used for continuous search spaces in the framework of evolution strategies [13, 14, 18] as well as for hybrid parallel versions of evolutionary algorithms and simulated annealing [10, 14] It was recognized several times that these fine grained parallel algorithms may be regarded as cellular automata [17, 19, 21] To provide a theoretical framework to study the differences we ....

.... but also offer a better solution quality than traditional GAs [11, 5, 17, 12, 16] Recently, this approach was also used for continuous search spaces in the framework of evolution strategies [13, 14, 18] as well as for hybrid parallel versions of evolutionary algorithms and simulated annealing [10, 14]. It was recognized several times that these fine grained parallel algorithms may be regarded as cellular automata [17, 19, 21] To provide a theoretical framework to study the differences we formally present a GA as a probabilistic cellular automaton, in which all genetic operators are applied ....

G. Rudolph. Massively parallel simulated annealing and its relation to evolutionary algorithms. Evolutionary Computation, 1(4):361--382, 1993.

No context found.

G. Rudolph. Massively Parallel Simulated Annealing and its Relation to Evolutionary Algorithms. Evolutionary Computation, pages 361--382, 1994.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC