Stefano Pizzuti

Urban traffic flow forecasting through statistical and neural network bagging ensemble hybrid modeling

In this paper we show a hybrid modeling approach which combines Artificial Neural Networks and a simple statistical approach in order to provide a one hour forecast of urban traffic flow rates. Experimentation has been carried out on three different classes of real streets and results show that the proposed approach outperforms the best of the methods it puts together.

Combining Back-Propagation and Genetic Algorithms to Train Neural Networks for Ambient Temperature Modeling in Italy

This paper presents a hybrid approach based on soft computing techniques in order to estimate ambient temperature for those places where such datum is not available. Indeed, we combine the Back-Propagation (BP) algorithm and the Simple Genetic Algorithm (GA) in order to effectively train neural networks in such a way that the BP algorithm initialises a few individuals of the GAs population. Experiments have been performed over all the available Italian places and results have shown a remarkable improvement in accuracy compared to the single and traditional methods.

Start-Up optimisation of a combined cycle power plant with multiobjective evolutionary algorithms

In this paper we present a study of the application of Evolutionary Computation methods to the optimisation of the start-up of a combined cycle power plant. We propose a multiobjective approach considering different objectives for the optimisation in order to reduce the pollution emissions and to maximise the efficiency of the plant. We compare a multiobjective evolutionary algorithm (NSGA-II) with 2 and 5 objectives on a software simulator and then we use different metrics to measure the performances. We show that NSGA-II algorithm is able to provide a set of solutions, defined as Pareto Front, that represent the best trade-off on the different objectives among those the decision maker can choose.

Evolving Complex Neural Networks

Complex networks like the scale-free model proposed by Barabasi-Albert are observed in many biological systems and the application of this topology to artificial neural network leads to interesting considerations. In this paper, we present a preliminary study on how to evolve neural networks with complex topologies. This approach is utilized in the problem of modeling a chemical process with the presence of unknown inputs (disturbance). The evolutionary algorithm we use considers an initial population of individuals with differents scale-free networks in the genotype and at the end of the algorithm we observe and analyze the topology of networks with the best performances. Experimentation on modeling a complex chemical process shows that performances of networks with complex topology are similar to the feed-forward ones but the analysis of the topology of the most performing networks leads to the conclusion that the distribution of input node information affects the network performance (modeling capability).

Evolving feed-forward neural networks through evolutionary mutation parameters

In this paper we show a preliminary work on evolutionary mutation parameters in order to understand whether it is possible or not to skip mutation parameters tuning. In particular, rather than considering mutation parameters as global environmental features, we regard them as endogenous features of the individuals by putting them directly in the genotype. In this way we let the optimal values emerge from the evolutionary process itself. As case study, we apply the proposed methodology to the training of feed-forward neural netwoks on nine classification benchmarks and compare it to other five well established techniques. Results show the effectiveness of the proposed appraoch to get very promising results passing over the boring task of off-line optimal parameters tuning.

Rotor Imbalance Detection in Gas Turbines Using Fuzzy Sets

The paper focuses on the application of fuzzy sets in fault detection. The objective is to detect faults to an industrial gas turbine, with emphasis on the imbalance occurred in the rotor of the gas turbine. Such a fault has a certain degree of uncertainty and an index based on fuzzy sets has been developed in order to provide a fault confidence degree (0 meaning no fault, 1 the fault has been detected by all the sensors). Experimentation has been carried out on three real industrial turbines and it has shown the reliability and effectiveness of the methodology.

On-line Identification of a Municipal Solid Waste Incinerator by Fully Tuned RBF Neural Networks

Abstract The paper describes an on-line identification algorithm to estimate the steam production of a municipal solid waste incinerator. The algorithm has to learn on-line the system dynamics due to the heavy disturbances acting on the incineration process. The learning algorithm is based on radial basis function networks and combines the growth criterion of the resource allocating network technique with an adaptive extended Kalman filter to update all the parameters of the networks.

Urban Traffic Flow Forecasting Using Neural-Statistic Hybrid Modeling

In this paper we show a hybrid modeling approach which combines Artificial Neural Networks and a simple statistical approach in order to provide a one hour forecast of urban traffic flow rates. Experimentation has been carried out on three different classes of real streets and results show that the proposed approach clearly outperforms the best of the methods it combines.

Fuzzy Optimization of Start-Up Operations for Combined Cycle Power Plants

In this paper we present a study on the application of fuzzy sets for the start-up optimisation of a combined cycle power plant. We fuzzyfy the output process variables and then we properly combine the resulting fuzzy sets in order to get a single value in the lattice [0,1] providing the effectiveness (zero bad, one excellent) of the given start-up regulations. We tested the methodology on a large artificial data set and we found an optimum which remarkably improves the solution given by the process experts.

ANALYSIS AND PREDICTION OF SPATIO-TEMPORAL FLAME DYNAMICS

In this paper we discuss novel methods of classification and prediction of spatio-temporal dynamics in extended systems. We tested these methods on simulated data for the KuramotoSivashinsky equation that describes unstable flame front propagation in uniform mixtures.