Uma solução heurística para o problema do carteiro chinês num grafo misto com aplicação à distribuição de bens e serviços públicos.

Abstract

In this work, variants of heuristic algorithms are established by Edmonds and Johnson and Frederickson, for the Chinese postman problem in a mixed graph. These variants can be applied to the problem of finding routes for distribution of public goods and services. As an application was used garbage collection of the city of Aracaju, Sergipe. The most important result of the work is that these variants perform better with respect to the application in computers than the Edmonds and Johnson and Frederickson originals, and that the worst case analysis shows that Frederickson's estimation, that is, the cost of a route found by the use of its algorithm is less than or equal to 5/3 of the cost of an optimal route, is also valid for the variant presented here. It can also be expected that in each case the studied variant presents a result with cost less than or equal to the result obtained by applying the original algorithms. This was verified in all the examples used during the development of this work.