Kamil Piętak

Agent-Based Simulation in AgE Framework

The chapter introduces AgE framework as a core for constructing agent based simulation systems. Its features are described against other solutions that may be used in the area of agent-based simulation. The discussion focuses on technical issues—the support for agent-specific services as well as the mechanisms allowing for extensibility and flexibility of the configuration of simulation models and systems. The considerations are illustrated by a simple case study, which aims at showing the differences between AgE and several selected tools for agent-based simulation.

Functional Integrity of Multi-agent Computational System Supported by Component-Based Implementation

In the paper a formalism is proposed to describe the hierarchy of multi-agent systems, particularly suitable for the design of a certain class of distributed computational intelligence systems. The notions of algorithms and dependencies among them are introduced, which allow for the formulation of functional integrity conditions for the whole system. General considerations are illustrated by modeling a specific case of an evolutionary multi-agent system. Component techniques introduced in AgE computing environment facilitate the implementation of the system in such a way that algorithm dependencies are represented as contracts, which support checking of the systems functional integrity.

Highly scalable Erlang framework for agent-based metaheuristic computing

Abstract Difficult search and optimization problems, usually solved by metaheuristics, are very often implemented in concurrent and parallel environment, as many metaheuristics (e.g. population- or agent-based) are inherently easy to parallelize. Therefore search for easy-to-use, robust and efficient frameworks dedicated for such computing methods, especially in the era of ubiquitous many and multi-core systems, is very desirable. Indeed, the development of multi-core architectures is incredibly fast and multicore CPUs can be found nowadays not only in supercomputers, but also in ordinary laptops or even phones. Efficient use of multicore architectures requires applying suitable languages and technologies, like Erlang. Its concurrency model, based on lightweight processes and asynchronous message-passing, seems very well suited for running massively concurrent code on many cores. Most of existing Erlang industrial applications are deployed on computers with up to 24 CPU cores, and there are hardly any reports on using Erlang on architectures exceeding 32 physical cores. In this paper we present our experiences with developing a concurrent Erlang-based computing platform, scaling computationally-intensive and memory-intensive applications up to 64 cores, using as examples global optimization and urban traffic planning problems.

Agent-Based Framework Facilitating Component-Based Implementation of Distributed Computational Intelligence Systems

The paper presents a framework particularly suitable for the design of a certain class of distributed computational intelligence systems based on the agent paradigm. A starting point constitutes a formalism utilizing the notions of algorithms and dependencies, which allows for the formulation of the system functional integrity conditions. Next, technological assumptions of AgE framework are presented and a direct mapping between the formalism and the implementation structure of the framework is discussed. The approach assumes that component techniques facilitate the realization of the particular system in such a way that algorithm dependencies are represented as contracts. These allow to support the verification of the system’s functional integrity. Selected technical aspects of the framework design illustrate the considerations of the paper.

Lightweight Distributed Component-Oriented Multi-Agent Simulation Platform.

Existing solutions for agent-based systems turn out to be limited in some applications, like agent-based computing or simulations, where very large numbers of clearly defined agents interact heavily within a closed system. In those cases, fully-fledged, FIPA1 compliant environment introduce unnecessary overhead, but simple tools fail to scale when confronted to bigger problems. In this paper, we introduce an alternative agent environment called AgE, targeted at medium-sized simulation and computational applications, which use multi-agent and computational intelligence paradigms, but does not need full FIPA compliancy, and would benefit from a component-based approach and distributed computing capabilities. After giving a short review of selected popular multi-agent platforms, the main features of AgE are presented. Next, some basic usability topics are addressed. Then the most interesting architectural aspects of the platform are discussed. Finally, AgE possibilities are demonstrated with two example applications. Keywords— agent-based computing, component-based systems, agent-based simulation

Multi-Domain Data Integration for Criminal Intelligence

The aim of the paper is to discuss the problem of heterogeneous data integration in LINK – a decision-support system for criminal analysis. In order to integrate and analyze various data sources, an object-based data model is proposed for each analyzed domain. From the technological side, this concept is supported by the component-oriented approach and tools (such as Eclipse RCP), which allow for flexible adding new domain objects. To verify the realized concept, a simple case study is given as an example of the integration results.

Distributed Agent-Based Platform for Large Scale Evolutionary Computations

In the paper a new volunteer computing environment dedicated for large-scale evolutionary computations is presented. It forms a distributed evolutionary multi-agent platform which utilizes Java applets as computational workers. After describing the general concept of the platform, the different possibilities of its deployment are discussed.

Toward hybrid platform for evolutionary computations of hard discrete problems

Abstract Memetic agent-based paradigm, which combines evolutionary computation and local search techniques in one of promising meta-heuristics for solving large and hard discrete problem such as Low Autocorrellation Binary Sequence (LABS) or optimal Golomb-ruler (OGR). In the paper as a follow-up of the previous research, a short concept of hybrid agent-based evolutionary systems platform, which spreads computations among CPU and GPU, is shortly introduced. The main part of the paper presents an efficient parallel GPU implementation of LABS local optimization strategy. As a means for comparison, speed-up between GPU implementation and CPU sequential and parallel versions are shown. This constitutes a promising step toward building hybrid platform that combines evolutionary meta-heuristics with highly efficient local optimization of chosen discrete problems.

Evolutionary Multi-Agent System in Planning of Marine Trajectories

The paper considers application of agent-based computing system, namely Evolutionary Multi-Agent System, to solving a difficult yet interesting problem of a marine glider path planning. Different version of mutations are compared both for EMAS and evolutionary algorithm parametrized in the most possibly similar manner to EMAS and the observed results show that the EMAS is better in most of the experiments.

Towards Multi-Agent Simulations Accelerated by GPU

At present, GPUs (Graphics Processing Units) are commonly used to speedup any kind of computations. In this paper we present how GPUs and Nvidia CUDA can be used to accelerate the updating of and agent state in Multi-Agent Simulations. We use the AgE (Agent Evolution) software framework written in Java, which supports agent-based computations. In our simulations agents represent living organisms that interact with the virtual habitat and with each other. At each step of the simulation thousands of agents update their state according to a defined set of rules. We use Java bindings for CUDA (JCUDA) to move massive computations to GPU.