Operation research and Meta heuristics

Operation research and Meta heuristics

Introduction

From the early years of the 80s, the investigation of combinatorial optimization problems is focused on the design of general strategies that serve to guide the heuristics. It has been called Meta heuristics. It is intelligently combine different techniques to explore the solution space (Siarry & Michalewicz, 2008). Although there are significant differences between the different methods developed so far, try to combine them all to a greater or lesser extent in the search intensified, selecting movements that will improve the assessment of the objective function, and diversification-accepting those other solutions even be worse through which one can evade the local optima.

The susceptible Meta heuristics grouped in various ways. Some classifications resort to successive changes of a solution to another in search of the optimal, while others make use of movements applied to an entire population of solutions (Basseur & Liefooghe, 2013). Employment, where appropriate, to guide memory space exploration of possible choices allows other grouping. A proposal for the classification of heuristics and meta heuristics used in combinatorial optimization is included in the figure, all having in common the need for solutions to initial changes to achieve better ones (Garcia, Chicano & Alba, 2008). Clearly there are many more techniques in this optimization time, but this classification can be a starting point for a better taxonomy of them.

Meta heuristics used in combinatorial optimization could be classified into three major groups. The first generalize the sequential search environment so that, once the process has been undertaken, a trajectory of a solution is another neighbor walking until it concludes. Processes acting on populations of solutions, evolving into generations of higher quality are included in the second group. The third is made up of artificial neural networks. This classification would be insufficient for those hybrid meta heuristics employing a greater or lesser extent, some strategies and other groups (Liang et.al, 2013). This event generates a desirable adaptive enrichment opportunities, where appropriate, to various combinatorial optimization problems.

Literature Review

In Artificial Intelligence is often called the label of any heuristic method that takes into account the use of problem-specific knowledge for resolution. In this sense (Reeves, 1995) defines a heuristic algorithm as a search procedure quasi-optimal solutions to a computational cost reasonable, without being able to guarantee the optimality of the solutions or determine how far away we find the optimal solution (Marti, Resende & Ribeiro, 2013). This research has contributed to the development of methods and procedures general strategies involving problem solving. Meta heuristics (Glover, 1986) are general strategies for intelligent design or improve heuristic procedures for solving problems with high performance (Melian, 2003). Simulated Annealing technique belongs to the family of meta heuristics of Generalized Randomized Hillclimbing (Zerovnik, 2003).

There are complex combinatorial optimization problems in various fields as economics, commerce, engineering, industry and medicine. However, often these problems are very difficult to solve in practice. Study this difficulty inherent to solve these problems belong in the field of the theory of computer science, and ...