International Journal on Artificial Intelligence Tools

As faults are unavoidable in large scale multiprocessor systems, it is important to be able to determine which units of the system are working and which are faulty. System-level diagnosis is a long-standing realistic approach to detect faults in multiprocessor systems. Diagnosis is based on the results of tests executed on the system units. In this work we evaluate the performance of evolutionary algorithms applied to the diagnosis problem. Experimental results are presented for both the traditional genetic algorithm (GA) and specialized versions of the GA. We then propose and evaluate specialized versions of Estimation of Distribution Algorithms (EDA) for system-level diagnosis: the compact GA and Population-Based Incremental Learning both with and without negative examples. The evaluation was performed using four metrics: the average number of generations needed to find the solution, the average fitness after up to 500 generations, the percentage of tests that got to the optimal solution and the average time until the solution was found. An analysis of experimental results shows that more sophisticated algorithms converge faster to the optimal solution.

In this paper, we introduce Lamarckian learning theory into the Clonal Selection Algorithm and propose a sort of Lamarckian Clonal Selection Algorithm, termed as LCSA. The major aim is to utilize effectively the information of each individual to reinforce the exploitation with the help of Lamarckian local search. Recombination operator and tournament selection operator are incorporated into LCSA to further enhance the ability of global exploration. We compare LCSA with the Clonal Selection Algorithm in solving twenty benchmark problems to evaluate the performance of LCSA. The results demonstrate that the Lamarckian local search makes LCSA more effective and efficient in solving numerical optimization problems.

This paper studies the problem of removing unused arguments from logic programs which have been constructed by a schema-based method. Our schema-based method constructs logic programs semi-automatically. These programs have clear structure which depicts the design decisions that have been taken for their construction. On the other hand, these programs have unused arguments. We propose a method that automatically removes the unused arguments from such programs. This method is based on fold-unfold transformations. We have developed a tool which implements our method and evaluates the initial and the transformed programs. This logic program transformation tool is presented as well.

Computing inference from a given knowledgebase is one of the key competences of computer science. Therefore, numerous formalisms and specialized inference routines have been introduced and implemented for this task. Typical examples are Bayesian networks, constraint systems or different kinds of logic. It is known today that these formalisms can be unified under a common algebraic roof called valuation algebra. Based on this system, generic inference algorithms for the processing of arbitrary valuation algebras can be defined. Researchers benefit from this high level of abstraction to address open problems independently of the underlying formalism. It is therefore all the more astonishing that this theory did not find its way into concrete software projects. Indeed, all modern programming languages for example provide generic sorting procedures, but generic inference algorithms are still mythical creatures. NENOK breaks a new ground and offers an extensive library of generic inference tools based on the valuation algebra framework. All methods are implemented as distributed algorithms that process local and remote knowledgebases in a transparent manner. Besides its main purpose as software library, NENOK also provides a sophisticated graphical user interface to inspect the inference process and the involved graphical structures. This can be used for educational purposes but also as a fast prototyping architecture for inference formalisms.

We argue that A*, the popular technique for path-finding for NPCs in games, suffers from three limitations that are pertinent to game worlds: (a) the grid maps often restrict the optimality of the paths, (b) A* paths exhibit wall-hugging behavior, and (c) optimal paths are more predictable. We present a new algorithm, VRA*, that varies map-resolution as needed, and repeatedly calls A*. We also present an extension of an existing post-smoothing technique, and show that these two techniques together produce more realistic looking paths than A*, that overcome the above limitations, while using significantly less memory and time than A*.

Recently, Question Answering has been a hot topic in the research of information retrieval. Question Classification plays a critical role in most Question Answering systems. In this paper, a new approach to classifying questions using Profile Hidden Markov Models (PHMMs) is proposed. The generalization strategies to extract the pattern instances of questions by selective substitution are discussed. Then the classification method with pattern instances' structural features is investigated. Experimental results show that the PHMM based question classifier can reach the accuracy of 92.2% and significantly outperforms most of the state-of-the-art systems.