In this chapter an algorithm is presented that incorporates many of the mechanisms for adaptivity described in Chapter 3. The algorithm began as DHPC and developed incrementally. Originally, the algorithm was to be used to investigate the performance of variants of DHPC and PPM. However, as the algorithm developed, it became clear that the algorithm’s flexibility and integration of diverse, interacting features was of interest in its own right. In this chapter, the algorithm, called the SAKDC83algorithm, is discussed in detail and the results of experiments that explore its parameter space are presented. The exploration of SAKDC’s parameter space not only lends experimental support to the theory presented in Chapter 3, but provides guidelines for practitioners working with Markov algorithms.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
- An Experimental Adaptive Algorithm
Ross N. Williams
- Springer US
- Chapter 4