Thomas Kurz

Simulation of a Self-Optimising Digital Ecosystem

Small and medium enterprises (SMEs) most often lack resources for custom-made software solutions that support best their core businesses. In this paper a simulation framework for self-optimising SME networks, namely Evolutionary Environment Simulator (EvESimulator), is introduced. The Evolutionary Environment, which the framework bases on, is an infrastructural component for distributed and decentralised service creation and service improvement based on mechanisms that are operating similarly in the living environment. This paper concentrates on the simulation of these mechanisms for SME networks. Built upon the widely-used simulation framework Repast, the cooperation behaviour of companies is investigated that use self-optimizing software services and distribute related information amongst them. The simulator is capable of importing real-world data from real businesses thereby enabling conceptual studies and hypothesis testing. It is applied in the context of three utilisation scenarios that investigate critical mass for sustainability, clustering in general, and usage-based clustering. The first results of the EvESimulator reveal a dynamic creation of a growing network of businesses that is clearly outperforming centralized topologies in the long run.

Model-transformation-based Software Generation utilizing Natural language notations

Adaptive software generation based on natural language notations is a key challenge in software industries. In the present work, the ability of stakeholders to formulate requirements that can directly be used in software production and software generation bases on a natural-language-based formal notation called Semantics for Business Vocabularies and Business Rules (SBVR), a recent OMG standard. SBVR-based model-to-model transformations are utilized to generate prototypical user interfacing code to demonstrate that non-IT business stakeholders can create framework-based software out of the box.

Multi-Agent simulation framework for interdisciplinary research in digital ecosystems

The conceptualization and ongoing implementation of a network and service infrastructure for SMEs that bases on biologically inspired similes for service mediation needs thorough simulation and emulation capabilities that accompany that development. Digital Ecosystems (DE) are demand-driven self-organising agent environments exhibiting complex behaviour during utilisation and growth. To optimise this infrastructure, the capability of modelling of central DE mechanisms is a key success factor. In this paper the fully distributed Evolutionary Environment Simulator (EvESim) is discussed that enriches the DE infrastructure by simulation and emulation capabilities for the benefit of interdisciplinary research.

Collaborative Software Engineering with a Digital Ecosystem

Adaptation to changing requirements is one of the predominant challenges in distributed development. Software projects exceeding certain scales cannot be addressed by small-sized companies because of the incapability of meeting the financial guarantees and also because of staffing problems. In spite of their advantage of being more flexible, small companies when teaming up are usually overextended due to the overhead of communication and synchronization costs. A need emerges here for an infrastructure endorsing decomposability of software specifications and semi-automatic re-composition of the implemented components to satisfy the original requirements. This infrastructure facilitates a more efficient risk management due to the more finely grained specification. Such an infrastructure is presented here as a result of an ongoing research and implementation activity including real-world testing spanning several European countries and regions

A distributed agent-based approach for interdisciplinary collaboration

The unpredicted growth and high user acceptance of social networks motivates the need for powerful and flexible modeling tools that can support analysis and further improvement of these platforms. In this paper we introduce an agent-based distributed modeling framework that is capable of simultaneously simulating and emulating user behaviour in social networks. This framework is independent of the underlying technology and can thus be utilized to study dynamic behaviour in a variety of settings. It constitutes a platform for interdisciplinary collaboration based on its technological conception, which is outlined in this paper together with a first proof-of-concept trial in context with a large social network in Southern America.

Emulation of social network ecosystems: Dynamic modeling and visualisation on the example of the Indian Agropedia project

Simulation of social networks that is based on abstract features of existing infrastructures might lack reflectiveness. This paper discusses the multi-agent EvESim framework and its application to study processes in the social network Agropedia, which is a digital knowledge repository with an open platform for learning and sharing information related to Indian agriculture. The emulation capability of the EvESim framework delivers insights into social network bootstrap and growth.

Knowledge Resources - A Knowledge Management Approach for Digital Ecosystems

The paper at hand presents an innovative approach for the conception and implementation of knowledge management in Digital Ecosystems. Based on a reflection of Digital Ecosystem research of the past years, an architecture is outlined which utilizes Knowledge Resources as the central and simplest entities of knowledge transfer. After the discussion of the related conception, the result of a first prototypical implementation is described that helps the transformation of implicit knowledge to explicit knowledge for wide use.