Artificial Intelligence in Engineering最新文献

Concern for environmental issues has increased in recent years. Waste production influences humanity's future. The alga, an ubiquitous single-celled plant, can thrive on industrial waste, to the detriment of water clarity and human activities. To avoid this, biologists need to isolate the chemical parameters of these rapid population fluctuations. This paper proposes a Fuzzy–Rough Estimator of Algae Populations (FuREAP), a hybrid system involving Fuzzy Set and Rough Set theories that estimates the size of algae populations given certain water characteristics. Through dimensionality reduction, FuREAP significantly reduces computer time and space requirements. Also, it decreases the cost of obtaining measurements and increases runtime efficiency, making the system more viable economically. By retaining only information required for the estimation task, FuREAP offers higher accuracy than conventional rule induction systems. Finally, FuREAP does not alter the domain semantics, making the distilled knowledge human-readable. The paper addresses the problem domain, architecture and modus operandi of FuREAP, and provides and discusses detailed experimental results.

In this paper, a new time-series predication method is proposed based on pattern analysis. In this method, basic patterns and their probabilities are extracted from a time series. A probabilistic relaxation method is employed to classify the probability vectors of the basic patterns. In order to verify the effectiveness of the proposed method, several experiments are carried out on a simulation signal and real data. The results show that the proposed method has advantages over existing methods in some applications.

This article reviews basic principles of animal learning and their potential contribution to the adaptation of user interfaces. The principles of classical conditioning, as well as a model that predicts most of the conditioning phenomena, are exposed. This paradigm has been widely studied in fields like Psychology, Biology and Computational Neuroscience, since the properties for stimuli association observed in experiments defined under this principle are important for the understanding of human and animal behavior. We present a direct application of these computational properties to the development of a certain kind of intelligent user interface. The main contribution is a general methodology for intelligent interfaces definition that can adapt themselves in an on-line fashion and without any a priori information of their interaction with the user. This adaptive paradigm outperforms conventional human–interface interaction, yielding more elaborated patterns of behavior where spatial and temporal associations among stimuli play an important role. The achieved upgrading is concerned with a significant effort: understanding user interfaces as living organisms, and identifying the set of stimuli and responses that determine the interaction with the user. Finally, the proposed paradigm is shown to successfully accomplish the adaptation of a customized interface in order to speed up its interaction with the user. The main differences with traditional sequence learning models are also discussed.

The control and prediction of wastewater treatment plants poses an important goal: to avoid breaking the environmental balance by always keeping the system in stable operating conditions. It is known that qualitative information — coming from microscopic examinations and subjective remarks — has a deep influence on the activated sludge process. In particular, on the total amount of effluent suspended solids, one of the measures of overall plant performance. The search for an input–output model of this variable and the prediction of sudden increases (bulking episodes) is thus a central concern to ensure the fulfillment of current discharge limitations. Unfortunately, the strong interrelation between variables, their heterogeneity and the very high amount of missing information makes the use of traditional techniques difficult, or even impossible. Through the combined use of several methods — rough set theory and artificial neural networks, mainly — reasonable prediction models are found, which also serve to show the different importance of variables and provide insight into the process dynamics.

The application of formal methods in software engineering and hardware design has become an important field of research. It aims at minimizing time to market and reduce the overall development costs. While formal verification, e.g. model-checking, is widely used, methods for helping programmers or engineers in locating and fixing faults within a hardware design or software are rarely available. In this paper we describe part of the advanced diagnosis and measurement selection capabilities of the model-based diagnosis tool VHDLDIAG designed for (semi)automatically locating bugs in VHDL programs. VHDL is an Ada-like and widely used hardware description language. VHDL programs are converted into logical descriptions which are then used by a diagnosis engine for detecting the parts of the program responsible for an observed misbehavior. The results of diagnosis, i.e. the malfunctioning program fragments, are mapped back to the program code. Because of the logical description used VHDLDIAG can be applied to a wide range of programs from small to very large ones with up to thousands of MBytes of source code. This paper presents techniques which use multiple versions of a design in diagnosis, as well as the measurement selection process used in VHDLDIAG. Formal definitions and performance results using real-world VHDL programs are given.

The richness of the constraint satisfaction problem (or CSP) in representing combinatorial search maladies has resulted in a torrent of techniques for efficiently solving them. These techniques have focused on discovering better backtrack points, learning from dead-ends and avoiding repetitious interference, problem reduction method and the use of network heuristics. Much of this research has derived innovative methods for solving the CSP, however, the evaluations of the techniques have remained diverse and in many cases, statistically inaccurate.

Another issue with regard to the performance measurement of constraint satisfaction techniques is the inability to model computational constraint processing cost. It is not uncommon to find evaluations that are based on CSPs that differ only on the percentage of constraints and the tightness of each constraint. This may be justifiable if it can be established that they are the only contributing factors of the performance variable. The three aspects mentioned above comprise this paper's main focus points. They come under the general headings of Modelling CSP Difficulty, Modelling Constraint Cost and Elucidating Major Performance Factors respectively. This paper seeks to provide a set of proposals with respect to the above three well-known areas so as collectively to enhance the robustness of evaluations conducted in the field of constraint satisfaction.

In this paper, we propose a genetic algorithm that generates and assesses assembly plans. An appropriately modified version of the well-known partially matched crossover, and purposely defined mutation operators allow the algorithm to produce near-optimal assembly plans starting from a randomly initialised population of (possibly non-feasible) assembly sequences. The quality of a feasible assembly sequence is evaluated based on the following three optimisation criteria: (i) minimising the orientation changes of the product; (ii) minimising the gripper replacements; and (iii) grouping technologically similar assembly operations. Two examples that endorse the soundness of our approach are also included.

This paper proposes a heuristic approach based on Hopfield model of neural networks to solve the problem of routing which constitutes one of the key aspects of the topological design of computer networks. Adaptive to changes in link costs and network topology, the proposed approach relies on the utilization of an energy function which simulates the objective function used in network optimization while respecting the constraints imposed by the network designers. This function must converge toward a solution which, if not the best is at least as close as possible to the optimum. The simulation results reveal that the end-to-end delay computed according to this neural network approach is usually better than those determined by the conventional routing heuristics, in the sense that our routing algorithm realizes a better trade-off between end-to-end delay and running time, and consequently gives a better performance than many other well-known optimal algorithms.