Frantisek Zboril

Towards Accelerated Computation of Atmospheric Equations Using CUDA

Main objective of this paper is to outline possibleways how to achieve a substantial acceleration in caseof advection-diffusion equation (A-DE) calculation,which is commonly used for a description of thepollutant behavior in atmosphere. A-DE is a kind ofpartial differential equation (PDE) and in general caseit is usually solved by numerical integration due to itshigh complexity. These types of calculations are timeconsuming thus the main idea of our work is to adoptCUDA platform and commodity GPU card to do thecalculations in a faster way. The solution is based onmethod of lines with 4th order Runge-Kutta scheme tohandle the integration. As a matter of fact, the selectedapproach involves number of auxiliary variables andthus the memory management is critical in order toachieve desired performance. We have implementedseveral possible solutions that use different memoryaccess schemes. Detailed evaluation is provided in thispaper where the obtained results show a tremendousprocessing speed up in comparison to CPU.

Intelligent Agent Platform and Control Language for Wireless Sensor Networks

In this paper we present a way how artificial agents can be brought into the world of wireless sensor networks. Architecture of agent platform which is implemented on individual nodes of such network is also introduced and its functionality on the conceptual level is described. Important part of the platform is a module for an agent control language interpretation. Such language is an original low level control language called Agent Low Level Language (ALLL). Other modules are responsible for data logging, agent migration and for general control of the platform execution cycle. There are also some additional computational services

Hierarchical Model of Trust in Contexts

Trust may be an important part of agent’s reasoning when resources, services or information are sought inside a multiagent system. Decision which agent to contact when more acceptable possibilities is naturally done with respect to expected outcome of forthcoming cooperation, so the partner with the best quality should be probably contacted first. Typically for one agent or device a trust may be evaluated in many contexts. Particular kinds of trust need not to be independent and trust in one aspect could influence trust in another aspect. We present a model of trusts as a multilevel computational graph where trusts are computed in conjunction on some trusts at preceding levels.

Event driven multi-context trust model

Agent reasoning in large scale multi-agent systems requires techniques which often work with uncertainty and probability. In our research, we use trust and reputation principles to support agent reasoning and decisioning. Information about agents past behaviour and their qualities are transformed to multi-context trust. It allows to view a single agent from different point of views, because agents are judged in different aspects — contexts. In this paper we describe event driven multi-context trust model as extension of Hierarchical Model of Trust in Contexts (HMTC), when different types of events causes trust updates. This extension of HMTC also provides some solutions for avoiding conflicts which may appear in previous HMTC.

Automata for Agent Low Level Language Interpretation

Agent platforms are usually strongly bounded with the hardware on which the agent works. Although the FIPA specifications declare what all should be part of such a platform, in some situation the agents are considered to be at bit lower level than the usual artificial agents are. In this text we present architecture of a system that works with an agent language that lies at quite low level of abstraction. The main part of this text is about automaton which accepts the Agent Low Level Language (ALLL) and also about the problem how this automaton can be implemented for the wireless sensor modules.

An Approach to ANFIS Performance

The paper deals with Adaptive neuro-fuzzy inference system (ANFIS) and its performance. Firstly, ANFIS is described as a hybrid system based on fuzzy logic/sets and artificial neural networks. Subsequently, modifications of ANFIS are proposed. The aim of these modifications is to improve performance, accuracy or reduce computational time. Finally, experiments are presented and findings are assessed.

GPU Accelerated Solver of Time-Dependent Air Pollutant Transport Equations

Main objective of this paper is to outline possible ways how to achieve a substantial acceleration in case of advection-diffusion equation (A-DE) calculation, which is commonly used for a description of the pollutant behavior in atmosphere. A-DE is a kind of partial differential equation (PDE) and in general case it is usually solved by numerical integration due to its high complexity. These types of calculations are time consuming thus the main idea of our work is to adopt compute unified device architecture (CUDA) software framework and commodity add-on card with generalpurpose graphics processing unit (GPU) to do the calculations in a faster way. The solution is based on method of lines with 4 th order Runge-Kutta scheme to handle the integration. As a matter of fact, the selected approach involves number of auxiliary variables and thus the memory management is critical in order to achieve desired performance. From a technical point of view, we have implemented a particular variant of the A-DE system, where the pollutant concentration is time-dependent. An efficient data handling is primarily based on the exploitation of shared memory blocks and texture caches inside GPU chip. Detailed evaluation of the obtained results is given in this paper where an astonishing execution speed up of GPU-based solution is demonstrated in comparison to standard CPU.

PNagent: A Framework for Modelling BDI Agents Using Object Oriented Petri Nets

Petri nets have proven to be a valuable tool for modelling concurrent systems. As such, they should be suitable for modelling agent and multi-agent systems, as these are inherently concurrent. Nevertheless, there are not many systems that allow modelling of the whole multi-agent system in Petri nets. The use of Petri nets is usually limited to sub-problems, such as verification of interaction protocols. This paper describes the main concepts of PNagent - a tool for modelling BDI agents using the formalism of object oriented Petri nets (OOPN). PNagent allows development, testing and running software agents in a consistent graphical environment. The framework is suitable for prototyping and experiments with both multi-agent applications and the particular agent architecture itself. At the same time, thanks to the formal nature of its underlying paradigm, it provides means for verification of agentpsilas properties.

Object Oriented Petri Nets Modelling Techniques Case Study

The importance of models used in the area of system design is growing. There were investigated and developed many methodologies of system design based on models - they are known as model-based design. These methodologies use semi-formal models and need for model transformations to check correctness or to get a final application. Formal models, by contrast, allow for clear modelling and checking correctness by simulation techniques as well as by formal verifications. The model is an executable program valid through all development stages including the target application. There are several suitable formalisms, one of them is a formalism of Petri nets, especially its object variant, object oriented Petri nets (OOPN). The paper deals with application of OOPN formalism in the area of system design and demonstrates modeling techniques of OOPN on the case study of conference system design.

Genetic Algorithm using Theory of Chaos

Abstract This paper is focused on genetic algorithm with chaotic crossover operator. We have performed some experiments to study possible use of chaos in simulated evolution. A novel genetic algorithm with chaotic optimization operation is proposed to optimization of multimodal functions. As the basis of a new crossing operator a simple equation involving chaos is used, concrete the logistic function. The logistic function is a simple one-parameter function of the second order that shows a chaotic behavior for some values of the parameter. Generally, solution of the logistic function has three areas of its behavior: convergent, periodic and chaotic. We have supposed that the convergent behavior leads to exploitation and the chaotic behavior aids to exploration. The periodic behavior is probably neutral and thus it is a negligible one. Results of our experiments confirm these expectations. A proposed genetic algorithm with chaotic crossover operator leads to more efficient computation in comparison with the traditional genetic algorithm.