Particular Domain Of Application Research Articles

A generic model for representing openness in multi-agent systems

Abstract Openness is a challenging property that may characterize multi-agent systems (MAS). It refers to their ability to deal with entities leaving and joining agent society over time. This property makes the MAS behaviour complex and difficult to study and analyze, hence the need for a representative model allowing its understanding. In this context, many models were defined in the literature and we propose to classify them into three categories: structural models, functional models and interactional models. The existing models were proposed either for representing structural openness or for modelling functional or interactional ones independently. But, none of them was oriented to represent MAS openness in a global way while considering its three aspects at once. Besides, each one was defined in order to realize a specific objective and in a particular domain of application. In this paper, we propose an evolving KAGR graph. The latter provides a common understanding of openness and unifies its structural, functional and interactional aspects in a generic way. Our model is finally tested and validated on a multi-agent rescue simulator.

Supervised Discovery of Unknown Unknowns through Test Sample Mining (Student Abstract)

Given a fixed hypothesis space, defined to model class structure in a particular domain of application, unknown unknowns (u.u.s) are data examples that form classes in the feature space whose structure is not represented in a trained model. Accordingly, this leads to incorrect class prediction with high confidence, which represents one of the major sources of blind spots in machine learning. Our method seeks to reduce the structural mismatch between the training model and that of the target space in a supervised way. We illuminate further structure through cross-validation on a modified training model, set up to mine and trap u.u.s in a marginal training class, created from examples of a random sample of the test set. Contrary to previous approaches, our method simplifies the solution, as it does not rely on budgeted queries to an Oracle whose outcomes inform adjustments to training. In addition, our empirically results exhibit consistent performance improvements over baselines, on both synthetic and real-world data sets.

Multilayer document model for semantic document management services

PurposeThe purpose of this paper is to identify the benefits of an approach in which document management systems (DMSs) are based on a formal and explicit document model, primarily in terms of facilitating domain-specific customization.Design/methodology/approachWithin this paper, a generic document model is proposed. The model consists of two layers. A general purpose layer, which represents common features of the documents, and a domain-specific layer, modeling properties particular to application domain. The general purpose layer is based on ISO 82045, providing high degree of interoperability with other systems developed with respect to this set of standard.FindingsSplitting document model into the layers enables DMSs to be tailored for each particular domain of application, depending on the general purpose layer. The existence of domain-specific layer allows documents to be interpreted differently in different domains of application.Practical implicationsIn order to enable customization of DMS for a particular domain, the implementation of domain-specific document layer is required. Also, the proposed model does not explicitly deal with document dynamics.Originality/valueThe proposed document ontology is general enough to provide the representation of documents not depending on a specific scope of application, yet flexible enough to enable extensions through which domain-specific document features can be expressed. The separation of document model enables development of core DMS offering services relying explicitly on the general purpose layer on one hand, as well as domain-specific customization of DMS on the other.

Non-destructive soil inspection using an efficient 3D software–hardware heat equation solver

Soil inspection by means of non-destructive evaluation (NDE) methods constitutes a subject of great research interest with practical implications beyond a particular domain of application. NDE methods often require the modelling of physical phenomena involving complex sets of equations whose computational solution has demanding requirements in terms of memory and computing power. One case of particular importance is that of the heat transfer equation, present in multiple situations. In this article, we will focus on NDE for soil inspection, with a particular application on the detection of shallowly buried antipersonnel mines in sandy soils. To this aim, a quasi-inverse procedure for the estimation of the depth of burial of the mine targets, and a full-inverse procedure for the characterization of the physical properties of other targets will be proposed. The efficient computational solution of these procedures is ensured via a software–hardware heat equation solver presented here.

Case-based reasoning and TRIZ: A coupling for innovative conception in Chemical Engineering

With the evolutions of the surrounding world market, researchers and engineers have to propose technical innovations. Nevertheless, Chemical Engineering community demonstrates a small interest for innovation compared to other engineering fields. In this paper, an approach to accelerate inventive preliminary design for Chemical Engineering is presented. This approach uses case-based reasoning (CBR) method to model, to capture, to store and to make available the knowledge deployed during design. CBR is a very interesting method coming from artificial intelligence, for routine design. Indeed, in CBR the main assumption is that a new problem of design can be solved with the help of past successful ones. Consequently, the problem solving process is based on past successful solutions, therefore, the design is accelerated but creativity is limited and not stimulated. Our approach is an extension of the CBR method from routine design to inventive design. One of the main drawbacks of this method is that it is restricted in one particular domain of application. To propose inventive solution, the level of abstraction for problem resolution must be increased. For this reason CBR is coupled with the TRIZ theory (Russian acronym for Theory of Solving Inventive Problem). TRIZ is a problem solving method that increases the ability to solve creative problems thanks to its capacity to give access to the best practices in all the technical domains. The proposed synergy between CBR and TRIZ combines the main advantages of CBR (ability to store and to reuse rapidly knowledge) and those of TRIZ (no trade-off during resolution, inventive solutions). Based on this synergy, a tool is developed and a mere example is treated.

Variable length representation in evolutionary electronics.

This work investigates the application of variable length representation (VLR) evolutionary algorithms (EAs) in the field of Evolutionary Electronics. We propose a number of VLR methodologies that can cope with the main issues of variable length evolutionary systems. These issues include the search for efficient ways of sampling a genome space with varying dimensionalities, the task of balancing accuracy and parsimony of the solutions, and the manipulation of non-coding segments. We compare the performance of three proposed VLR approaches to sample the genome space: Increasing Length Genotypes, Oscillating Length Genotypes, and Uniformly Distributed Initial Population strategies. The advantages of reusing genetic material to replace non-coding segments are also emphasized in this work. It is shown, through examples in both analog and digital electronics, that the variable length genotype's representation is natural to this particular domain of application. A brief discussion on biological genome evolution is also provided.

Reciprocal Variable Feedback: Induced Sensing for Nonlinear Systems Design and Control

System performance can be significantly improved when both the design of the plant and of the controller are considered concurrently. Control theory can be applied to a broad variety of systems, including those that are physical in nature and many that are not. Despite the generality of control theory, there are many situations in which opportunities are missed for using less conservative control laws and simpler overall implementations. This is due to the use of formulations that do not explicitly reveal the existence of intrinsic information pertaining to the particular domain of application. Such is the case with many physical systems. However, the various constraints associated with physical reality (in the form of principles, laws, etc.) open up several possibilities which can be exploited for system design and control. In this paper, we propose the Reciprocal Variable Feedback principle as a means for facilitating the control of plants with complicated nonlinear dynamics in the presence of parameter and/or structural uncertainty. The RVF principle exploits the effort-flow relationships associated with power interactions in order to assist in the design and control of physical processes. This is accomplished by using appropriate sensors instead of computation based on models (e.g., feedback linearization) and can be implemented within many physical domains. A motion control example is used to provide insight into the nature of the principle. It is expected that in the future, additional principles will be identified and introduced for integrating design with the control of dynamical systems.

Heuristic scheduling of the DRS communication system

The DRS (Data Relay Satellite) System is a European Space Agency communication system whose management is particularly complex. Nevertheless, the particular domain of application is an example of problems that are becoming quite common in the communication scenarios enabled by current technology. This paper describes a system named DRS-Sched, a knowledge-based scheduler aimed at solving an important aspect of the whole DRS scheduling problem. In particular, the problem model addressed is rather distinctive: the activities to be scheduled have a temporal flexibility over both their allocation and duration intervals, and temporal preferenve functions over the allocation interval. Problem-solving in DRS-Sched integrates domain-specific predictive scheduling heuristics, that build a solution using a least commitment approach, and reactive heuristics, that try to improve the current schedule by locally modifying previous choices when a particular class of conflicts arises. The paper describes the problems addressed, and the basic algorithm and heuristics used in the current release of the system. An experimental evaluation is also presented, showing how the system satisfies quality requirements connected to different relevant aspects of the project.