Forecasting of the unknown end-of-life tire flow for control and decision making in urban solid waste management: A case study.

Efficient urban planning requires managers' experience and knowledge of reverse logistics in solid urban waste processes. Forecasting tools are needed to control, select and manage municipal solid waste. This paper presents the application of dynamic modeling approaches, namely, a linear autoregressive seasonal model, a model based on a FeedForward Artificial Neural Network and a Recurrent Neural Networks model, in order to forecast the unknown flows of end-of-life tires 12 months ahead. The models were identified using a database comprising four years of historical series related to the unknown flows of end-of-life tires. These were obtained through an exploratory analysis based on the annual sales reports of new tires issued by the Brazilian Institute of Geography and Statistics and reports related to the number of vehicles in circulation issued by Brazil's National Traffic Department. The results show that the models are able to carry out consistent forecasts over the horizon of a year ahead and the predictions are capable of identifying seasonalities and supporting decision making in urban waste management.

A hybrid clustering approach for multivariate time series - A case study applied to failure analysis in a gas turbine.

A clustering problem involving multivariate time series (MTS) requires the selection of similarity metrics. This paper shows the limitations of the PCA similarity factor (SPCA) as a single metric in nonlinear problems where there are differences in magnitude of the same process variables due to expected changes in operation conditions. A novel method for clustering MTS based on a combination between SPCA and the average-based Euclidean distance (AED) within a fuzzy clustering approach is proposed. Case studies involving either simulated or real industrial data collected from a large scale gas turbine are used to illustrate that the hybrid approach enhances the ability to recognize normal and fault operating patterns. This paper also proposes an oversampling procedure to create synthetic multivariate time series that can be useful in commonly occurring situations involving unbalanced data sets.