Maintenance optimization under uncertainties using interval methods & evolutionary strategies

Abstract

In this paper, the authors present two examples of maintenance optimization under uncertainties using interval arithmetic methods and evolutionary strategies. In the first example, they show the use of Interval Arithmetic to solve a single variable global optimization problem related to a replacement at predetermined age for a single component. Interval arithmetic considers the uncertainty of all the parameters at the same time and is able to provide strict bounds for the optimum solution. However, for complex problem, the amount of time required does pose a limitation. In the second example, they consider a multi-component system to determine the range of coordinated maintenance frequencies under to uncertainties. The example presented confirm that the proposed approach based on evolutionary strategies can be used to determine the range of an optimization model with uncertainties on the parameters as it produces a very good approximation. The approach is faster than other approaches and requires no assumption about the function to be optimized.