Abstract
A priori knowledge of the behavior of the aggregates is helpful for controlling the course of the algorithm such that chances for a successful run increase. However this behavior depends on the instance. It is therefore interesting to characterize instances with a number of parameters. These parameters do not have to be known at the start of the algorithm. They can be estimated ‘on the fly’ as long as these estimates have become accurate enough when they are used for deriving decisions. A study of the generic behavior of the aggregates is therefore useful. This behavior depends on the distribution of scores over the states. This distribution is not known in general, and consequently, we have to rely on assumptions and experiments to perform the kind of analysis necessary for discovering the parameters characterizing an instance. For the annealing applications that we have encountered the observations and conclusions of this chapter seem to be in excellent correspondence. However, applying these in implementations has to be done with care and in combination with generally valid criteria. The next chapter will take this task as its subject.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1989 Kluwer Academic Publishers
About this chapter
Cite this chapter
Otten, R.H.J.M., van Ginneken, L.P.P.P. (1989). Score Densities. In: The Annealing Algorithm. The Kluwer International Series in Engineering and Computer Science, vol 72. Springer, Boston, MA. https://doi.org/10.1007/978-1-4613-1627-5_7
Download citation
DOI: https://doi.org/10.1007/978-1-4613-1627-5_7
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4612-8899-2
Online ISBN: 978-1-4613-1627-5
eBook Packages: Springer Book Archive