International Journal on Artificial Intelligence Tools

The least square method is in generally used for curve fitting problems. We here propose a fuzzy S-curve regression model to deal with the case in which the observed data are given by fuzzy numbers. The fuzzy regression curve, obtained for project control and predicting the progress of large-scale or small-scale engineering, is smoothly connected by a Takagi-Sugeno (T-S) fuzzy model. This paper also proposes the concept that the upper bound and lower bound are given instead of the confidence interval when the observed data are not obtained exactly. Based on the project cash flow and progress payment records of an example project taken from the Department of Rapid Transit Systems, Taipei City Government, this model is demonstrated and tentative conclusions concerning the model are given. The S-curve equation developed here could be used in a variety of applications related to project control for the management of working capital for construction firms.

We investigate the circuit complexity of classification problems in a machine learning setting, i.e. we attempt to find some rule that allows us to calculate a priori the number of threshold gates that is sufficient to achieve a small error rate after training a circuit on sample data . The particular threshold gates are computed by a combination of the classical perceptron algorithm with a specific type of stochastic local search. The circuit complexity is analysed for depth-two and depth-four threshold circuits, where we introduce a novel approach to compute depth-four circuits. For the problems from the UCI Machine Learning Repository we selected and investigated, we obtain approximately the same size of depth-two and depth-four circuits for the best classification rates on test samples, where the rates differ only marginally for the two types of circuits. Based on classical results from threshold circuit theory and our experimental observations on problems that are not linearly separable, we suggest an upper bound of threshold gates as sufficient for a small error rate, where .

This paper proposes a fuzzy Lyapunov method for stability analysis of nonlinear systems represented by Tagagi-Sugeno (T-S) fuzzy model. The fuzzy Lyapunov function is defined in fuzzy blending quadratic Lyapunov functions. Based on fuzzy Lyapunov functions, some stability conditions are derived to ensure nonlinear systems are asymptotic stable. By using parallel distributed compensation (PDC) scheme, we design a nonlinear fuzzy controller for the nonlinear system. This control problem will be reformulated into linear matrix inequalities (LMI) problem.

This paper deals with the need for generic software development tools in evolutionary computations (EC). These tools will be essential for the next generation of evolutionary algorithms where application designers and researchers will need to mix different combinations of traditional EC (e.g. genetic algorithms, genetic programming, evolutionary strategies, etc.), or to create new variations of these EC, in order to solve complex real world problems. Six basic principles are proposed to guide the development of such tools. These principles are then used to evaluate six freely available, widely used EC software tools. Finally, the design of Open BEAGLE, the framework developed by the authors, is presented in more detail.

We present MABLE, a fully implemented programming language for multiagent systems, which is intended to support the automatic verification of such systems via model checking. In addition to the conventional constructs of imperative programming languages, MABLE provides a number of agent-oriented development features. First, agents in MABLE are endowed with a BDI-like mental state: they have data structures corresponding to beliefs, desires, and intentions, and these mental states may be arbitrarily nested. Second, agents in MABLE communicate via ACL-like performatives: however, neither the performatives nor their semantics are hardwired into the language. It is possible to define the performatives and the semantics of these performatives independently of the system in which they are used. Using this feature, a developer can explore the design space of ACL performatives and semantics without changing the target system. Finally, MABLE supports automatic verification via model checking. Claims about the behaviour of a MABLE system can be expressed in a linear-time BDI-like logic, and the truth, or otherwise, of these claims can be automatically determined. Following a description of the MABLE language and the language of MABLE claims, we present two case studies to illustrate the language and its use in the verification of multiagent systems. We then describe the key ideas underpinning the current implementation of MABLE. Finally, we survey related work, and discuss some avenues for future research.

Fuzzy association rules can reveal useful dependencies and interactions hidden in large gene expression data sets. However their derivation perplexes very difficult combinatorial problems that depend heavily on the size of these sets. The paper follows a divide and conquer approach to the problem that obtains computationally manageable solutions. Initially we cluster genes that more probably are associated. Thereafter, the fuzzy association rule extraction algorithms confront many but significantly reduced computationally problems that usually can be processed very fast.

The clustering phase is accomplished by means of an approach based on mutual information (MI). This approach uses the mutual information as a similarity measure. However, the numerical evaluation of the MI is subtle. We experiment with the main methods and we compare between them. As the device that implements the mutual information clustering we use a SOM (Self-Organized Map) based approach that is capable of effectively incorporating supervised bias.

After the mutual information clustering phase the fuzzy association rules are extracted locally on a per cluster basis. The paper presents an application of the techniques for mining the gene expression data. However, the presented techniques can easily be adapted and can be fruitful for intelligent exploration of any other similar data set as well.

In this paper we present the agent architecture development environment ADE, intended for the design, implementation, and testing of distributed agent architectures. After a short review of architecture development tools, we discuss ADE's unique features that place it in the intersection of multi-agent systems and development kits for single agent architectures. A detailed discussion of the general properties of ADE, its implementation philosophy, and its user interface is followed by examples from virtual and robotic domains that illustrate how ADE can be used for designing, implementing, testing, and running agent architectures.

The paper demonstrates the efficient use of hybrid intelligent systems for solving the classification problem of splice-junction gene sequences. The aim of the study is to obtain classification schemes able to recognize, given a sequence of DNA, the boundaries between exons and introns. Previous attempts to form efficient classifiers for the same problem using intelligent or standard statistical techniques are discussed throughout the paper. The authors propose the use of evolutionary neural logic networks, an advantageous approach for their ability to interpret their structure into expert rules, a desirable feature for field experts. Evolutionary neural logic networks in fact consist an innovative hybrid intelligent methodology, by which evolutionary programming techniques are used for obtaining the best possible topology of a neural logic network. The genetic programming process is guided using a context-free grammar and indirect encoding of the neural logic networks into the genetic programming individuals. Indicative classification results are presented and discussed in detail in terms of both, classification accuracy and solution interpretability.

The topological structures and robustness of the protein-protein interaction networks (PINs) is studied via random graph theory. Several global topological parameters are used to characterize PINs for seven species. Good evidence by correlation analysis support the fact that the seven PINs are well approximated by scale-free networks. It is also found that two of the species' PINs are well described by the hierarchical models. In particular, it is determined that the E. coli and the yeast PINs are well represented by the stochastic and deterministic hierarchical network models respectively. These results suggest that the hierarchical network model is a better description for certain species' PINs. Furthermore, it is demonstrated that PINs are robust with respective to failure, attack, random rewiring and edge deletion perturbations.

A metadata search approach is proposed to provide practical solutions to the natural language database query problem; where the metadata grow in largely linear manner and the search is linguistics-free. A new class of reference dictionary integrates four types of enterprise metadata: enterprise information models, database values, user-words, and query cases. The layered and scalable information models allow user-words to stay in original forms as users articulated them, as opposed to relying on permutations of individual words contained in the original query. A graphical representation method turns the dictionary into searchable graphs representing all possible interpretations of the input. A branch-and-bound algorithm then identifies optimal interpretations, which lead to SQL implementation of the original queries. Query cases enhance both the metadata and the search of metadata, as well as providing case-based reasoning to directly answer the queries. This design assures feasible solutions at the termination of the search, even when the search is incomplete (i.e., the results contain the correct answer to the original query). The necessary condition is that the text input contains at least one entry in the reference dictionary. The sufficient condition is that the text input contains a set of entries corresponding to a complete, correct single SQL query. Laboratory testing shows that the system obtained accurate results for most cases that satisfied only the necessary condition.