Solutions to Parallel and Distributed Computing Problems: Lessons from Biological Sciences

3.2 DESIGN ISSUES

Hybridization of heuristics involves a few major issues that can be classified as design and implementation, respectively. The former category concerns the hybrid algorithm itself, involving issues such as the functionality and architecture of the algorithm. The implementation consideration includes the hardware platform and environment on which the algorithm is to run. In this chapter, we show the difference between the design issues used to introduce hybridization and the implementation issues that depend on the execution model of the algorithms.

The taxonomy is kept as small as possible by proceeding in a hierarchical fashion as long as possible, but some choices of characteristics can be made independently of previous design choices, and thus will he specified as a set of descriptors from which a subset can be chosen.

3.2.1 Hierarchical Classification

The structure of the hierarchical portion of the taxonomy is shown in Fig. 3.1. A discussion of the hierarchical portion then follows.

3.2.1.1 Low Level versus High Level

At the first level, we can distinguish between low-level and high-level hybridizations.

Low-level hybridization addresses the functional composition of a single optimization method. In this hybrid class, a given function of a metaheuristic is replaced by another metaheuristic. In high-level hybrid algorithms, the different metaheuristics are self-contained. We include no direct relationship to the internal workings of a metaheuristic.

3.2.1.2 ...