Search

Among the techniques for solving complex optimization problems are heuristics and metaheuristics. This work presents three Metaheuristics: Ant Colony, Algorithm Genetic and Evolutionary Programming, which aims to establish a comparison of features, advantages and disadvantages for those Metaheuristics. A review of the literature is made, and also the design of the three algorithms in R program, and they are applied to a confection company, evaluating the performance in terms of its parameters, time calculation and quality solutions. Experimental results show that the Ant Colony algorithm shows adequate solutions, as well as it has more computational speed, when it is compared with the other two procedures designed in software R. In addition, with this, it was possible to define the best policy inventory of the company under study. [ABSTRACT FROM AUTHOR]

The Chinese Postman Problem is an optimization problem that aims to cover all the arcs of a graph, minimizing the total distance traveled. Can be applied to non-directed graphs (two-way streets), directed (one-way streets) or mixed (some two-way streets and other one-way). These arch for the route is done by algorithms that generate approximate solutions. In this work we used a genetic algorithm for the construction of routes that approximate the optimal solution. The main objective is to minimize the cost of the course of collection and transportation solid waste in the city of Irati (PR), Brazil. The application of the algorithm was performed in a downtown area. Were used formapping, real data courtesy of City Hall officials. With the help of the online application Google Earth, we obtained the geographical coordinates of the vertices of the graph associated with the problem. The results generated by the genetic algorithm were compared with the optimal solution obtained by LINGO ® 12.0 software. These results were satisfactory for this mall instance of problems analyzed. [ABSTRACT FROM AUTHOR]

Processes of a real world manufacturer involve a number of batch-processing operations. Generally, these operations require the decision on batch sizes of products and there is no specific rule to set batch sizes of all products simultaneously so that the sum of the weighted expected cycle times of all products is minimized. The weights represent different levels of importance for products. In this study, we proposed an efficient genetic algorithm to determine the batch size combination that minimizes this sum. The proposed GA was tested on three different data sets, having number of products ranging from 7 to 12. The results show that, for some cases, the proposed GA is effective in locating the global optimum solutions. [ABSTRACT FROM AUTHOR]