metaheuristic

(redirected from Meta heuristics)

metaheuristic

(algorithm, complexity, computability)
A top-level general strategy which guides other heuristics to search for feasible solutions in domains where the task is hard.

Metaheuristics have been most generally applied to problems classified as NP-Hard or NP-Complete by the theory of computational complexity. However, metaheuristics would also be applied to other combinatorial optimisation problems for which it is known that a polynomial-time solution exists but is not practical.

Examples of metaheuristics are Tabu Search, simulated annealing, genetic algorithms and memetic algorithms.
Mentioned in ?
References in periodicals archive ?
Even though the meta heuristics will work effectively while using preprocessed data, they also provide considerably good results even without data preprocessing.
The remainder of this paper is structured as follows; section II provides the review of literature, section III provides an overview of the available data mining techniques, section IV discusses the importance of meta heuristics for data mining, section V presents a brief overview of the concept of error existing in evolutionary algorithms, section VI discusses the concept of PSO, section VII discusses the mode of usage and advantages in using PSO for classification, section VIII compares PSO with other meta heuristic approaches, section IX discusses PSO's flexibility towards hybridization, section X presents the results and discussion and section XI concludes the study.
Hence meta heuristics are recommended for applications that can tolerate a small error rate (which encompasses a very large part of the available applications).
The speed and accuracy of a meta heuristic algorithm can be boosted up by incorporating the parallelization capability and utilizing the multi core CPUs and the many core GPUs, that have become a de facto in the current hardware specifications.
It has been discussed earlier that meta heuristic evolutionary algorithms provide near optimal solutions at the stipulated time.
On close inspection, it can be found that all the meta heuristic evolutionary algorithms have a component or mechanism built into it that introduces the much needed error into the system.
Hence analysis is to be carried out in terms of performance to determine their applicability in case of meta heuristic evolutionary algorithms.
Interest in meta-heuristics has generated the development of hybrid approaches [8] and recent significant advances have combined meta heuristics with other problem solving paradigms and improved their use in important application areas [23].
One reason is that many developers are afraid of dealing with complex multiple vehicle routing algorithms as this requires in depth experience with meta heuristics, which is of course not the primary focus of modern software engineers," Vincent Mayer says, (DNA's responsible engineer for application integration).
Topics include linear, nonlinear, queuing, decision analysis, meta heuristics, robust optimization, and large-scale networks.