In the world of liberalized power markets traditional power management concepts have come to their limits. Optimal pricing can no longer be achieved, e.g. for very short-time needs across grids. Power line overload and grid stability, increasingly resulting in regional or even global black-outs, are at stake. With the highly desirable expansion of renewable energy production these challenges are experienced in quite an amplified way: We argue that for this emergent technology the traditional top-down and long-term power management is obsolete, due to the wide dispersion and high unpredictability of wind and solar-based power facilities. In the DECENT0F 1 R&D initiative we developed a multi-level, bottom-up solution where autonomous collaborative software agents negotiate available energy quantities and needs on behalf of consumer and producer groups (the DEZENT algorithm). We operate within very short time intervals of assumedly constant demand and supply, in our case periods of 0.5sec (switching delay for a light bulb). The solution has proven to be secure against a relevant variety of malicious attacks. Within this time interval we are also able to manage the coordinated power distribution, and achieve grid stability. In this paper the main contribution is to make the negotiation strategies themselves adaptive across periods: We derive the dynamic distributed learning algorithm DECOLEARN from Reinforcement Learning principles for providing the agents with collaborative intelligence and at the same time proving substantially superior to conventional (static) procedures. We report briefly on our extensive comparative simulation experiments. Keywords: Distributed Energy Management, Reinforcement Learning, Multi-agent Systems, Smart Grids

Traffic in densely populated regions increasingly suffers from congestion problems, to an extent which e.g. substantially affects predictable transport planning. Due to the highly dynamic character of congestion forming and dissolving, no static approach like shortest path finding, applied globally or individually in car navigators, is adequate here. In this paper we outline our current work on a completely decentralized multi-agent bottom-up approach (termed BeeJamAJ) on multiple layers where car routing is handled through algorithms derived from honey bee behavior.

Both the coordination of international energy transfer and the integration of a rapidly growing number of decentralized energy resources (DER) throughout most countries cause novel problems for avoiding voltage band violations and line overloads. Traditional approaches are typically based on global off-line scheduling under globally available information and rely on iterative procedures that can guarantee neither convergence nor execution time. In this paper, we focus on operational limitation problems in power grids based on widely dispersed (renewable) energy sources. We introduce an extension to the DEZENT algorithm, a multi-agent based coordination system for DER, that allows for the feasibility verification in constant and predetermined time. We give a numerical example showing the legitimacy of our approach and mention ongoing and future work regarding its implementation and utilization.

The liberalization of electricity markets and the integration of high capacity unpredictable renewable resources (e.g. wind power) yield a higher utilization of transmission networks. The capacity of transmission networks can be increased by installing Power Flow Controllers (PFCs). When several PFCs with mutual impact are installed in different control areas a distributed coordination is needed. This paper presents a novel approach for a decentralized coordination of PFCs based on multi agent systems. Serial network devices are equipped with agents which submit messages about local state information along the system topology. Controlling agents installed at each PFC analyze the received messages to gather information about the current system topology and the sensitivity for control actions on network devices. Each controlling agent evaluates this information by use of a weighting function in order to avoid wrong control actions in case of conflicting control requests. Finally this paper shows a principle simulation scenario to illustrate the operation of the proposed multi agent control.

For regenerative electric power the traditional top- down and long-term power management is obsolete, due to the wide dispersion and high unpredictability of wind and solar based power facilities. In the R&D DEZENT1 project we developed a multi-level bottom- up solution where autonomous software agents negotiate available energy quantities and needs on behalf of consumers and producer groups. We operate within very short time intervals of assumedly constant demand and supply, in our case 0.5 sec (switching delay for a light bulb). We prove security against a relevant variety of malicious attacks. In this paper the main contribution is to make the negotiation strategies themselves adaptive across periods. We adapted a reinforcement Learning approach for defining and discussing learning strategies for collaborative autonomous agents that are clearly superior to previous (static) procedures. We report briefly on extensive comparative simulation.

Commercial transport planning as well as individual intra-city or inter-city traffic in densely populated regions, both in Europe and the US, increasingly suffer from congestion problems, to an extent which e.g. affects predictable transport planning substantially (except - so far - for overnight tours). Due to the highly dynamic character of congestion forming and dissolving, no static approach like shortest path finding, applied globally or individually in car navigators, is adequate here: Its use even makes things worse as can be frequently observed. In this paper we present a completely decentralized multi-agent approach (termed BeeJamA) on multiple layers where car or truck routing are handled through algorithms adapted from the BeeHive algorithms which in turn have been derived from honey bee behavior. We report on extensive distributed simulation experiments in the BeeJamA project which demonstrate a very substantial improvement over traditional congestion handling.

A world-wide trend towards renewable and ecologically clean forms of energy has been steadily growing. Private investments are encouraged and heavily subsidized in most of the European countries, through tax deductions, and even more through a very favorable refund program for feeding electric power from renewable sources into the public network. Due to the limited predictability of the output of renewable power capacities it has long become the policy of grid operators and large power distributors to cover the differences between demand and supply with immense reserve and balancing power capacities based on fossil, and thus predictable, energy sources. With growing renewable power feed-in the demand for reserve and balancing power grows over-proportionally. In 2005 the European Union for the Coordination of Transmission of Electricity (UCTE) demanded to impose an obligation on grid operators to reduce integration costs for renewable energy capacities. This could obviously be possible once the renewable capacities sources could serve as reserve capacity. Since these are widely distributed and dispersed, their combined effect may well be used to guarantee a stable supply. The remaining problem behind is that the largely unpredictable character of wind and solar power supply is to be administered financially and in terms of timely transmission. We introduce a novel solution for the distributed negotiation process, which is compatible with electric distribution procedures. This is part of our DEZENT (decentralized management of electric power distribution) project.

For a just-in-time production scheme in large manufacturing systems we describe a considerable methodological extension into an integrated automated manufacturing system that is supported by a distributed real-time computer system. This allows for higher flexibility in demand fluctuation as well as for a larger variety of products. At the same time better adaptability of transportation planning and transportation cost reduction are achieved through their integration. For this integrated scheduling we will present new bidding algorithms. Eventually, in an extension of the presented scheduling algorithms, novel fault tolerant strategies are included for overcoming or neutralizing the effect of transient machine failures. These are handled by cooperating local schedulers and managed to guarantee a minimal damage of schedules and its propagation, in the presence of failures. Through their integration into the production and transportation scheduling the advantages of the just-in-time approach (no storage costs) are preserved in principle while a near-optimal way for the affected jobs is found to meet their deadlines. The results of our extensive simulation experiments for a real production scenario are also discussed.

Establishing clean or renewable energy sources involves the problem of adequate management for the networked power sources, in particular since producers are at the same time also consumers, and vice versa. We describe the first phases of the joint R&D project DEZENT between the School of Computer Science and the College of Electrical Engineering at the University of Dortmund, devoted to decentralized and adaptive electric power management through a distributed real-time multiagent architecture. Unpredictable consumer requests or producer problems, under distributed control or local autonomy will be the major novelty. We present a distributed real-time negotiation algorithm involving agents on different levels of negotiation, on behalf of producers and consumers of electric energy. Despite the lack of global overview we are able to prove that in our model no coalition of malicious users could take advantage of extreme situations like arising from an abundance as much as from any (artificial) shortage of electric power that are typical problems in "free" or deregulated markets. Our multi-agent system exhibits a very high robustness against power failures compared to centrally controlled architectures. In extensive experiments we demonstrate how, in realistic settings of the German power system structure, the novel algorithms can cope with unforeseen needs and production specifics in a very flexible and adaptive way, taking care of most of the potentially hard deadlines already on the local group level (corresponding to a small subdivision). We further demonstrate that under our decentralized approach customers pay less than under any conventional (global) management policy or structure.

Establishing clean or renewable energy sources involves the problem of adequate management for the networked power sources, in particular since producers are at the same time also consumers, and vice versa. We describe the first phases of the joint R&D project DEZENT between the School of Computer Science and the College of Electrical Engineering at the University of Dortmund, devoted to decentralized and adaptive electric power management through a distributed real-time multiagent architecture. Unpredictable consumer requests or producer problems, under distributed control or local autonomy will be the major novelty. We present a distributed real-time negotiation algorithm involving agents on different levels of negotiation, on behalf of producers and consumers of electric energy. Despite the lack of global overview we are able to prove that in our model no coalition of malicious users could take advantage of extreme situations like arising from an abundance as much as from any (artificial) shortage of electric power that are typical problems in “free” or deregulated markets. Our multi-agent system exhibits a very high robustness against power failures compared to centrally controlled architectures. In extensive experiments we demonstrate how, in realistic settings of the German power system structure, the novel algorithms can cope with unforeseen needs and production specifics in a very flexible and adaptive way, taking care of most of the potentially hard deadlines already on the local group level (corresponding to a small subdivision). We further demonstrate that under our decentralized approach customers pay less than under any conventional (global) management policy or structure.

In this paper we present an energy efficient routing algorithm, BeeAdHoc, which is inspired from foraging principles of honey bees. The bee behavior was instrumental in designing efficient mobile agents, scouts and foragers, for routing in mobile ad-hoc networks. We did extensive simulations to verify that BeeAdHoc consumes significantly less wireless network card energy as compared to DSR, AODV, and DSDV, which are existing state-of-the-art routing algorithms, but without compromising traditional performance metrics, packet delivery ratio and delay.

We present I–Systems as a formal constraint-based approach for modeling and analyzing both autonomous and reactive behavior in a distributed system. Essentially it is a formalism of interacting finite automata. We demonstrate its incremental potential by stepwise modeling a solution for a synchronous communication problem.

Human Centered Design suffers from the same ignominy of Machine Centered Design that it promised to eradicate in the first place. This observation provided to us the grist for the mill in exploring the idea of Nature Centered Design. This design philosophy emphasizes the production of nature-like objects that conflate into the natural ecological system with an acceptable disorder in its entropy. This is only possible if thaumaturgy of Nature is used to cope with awe as a driving impulse to the awareness and consciousness of the designer. He can then explore the laws of Nature and apply them in the design of technological systems that are natural to interact with. In this paper we do a per-vestigation of the natural ecological system with a sentient contemplation to articulate the architecture of an information system based on the ecological relationships of natural objects in Nature.

The need for supporting CSCW applications with heterogeneous and varying user requirements call for adaptive and reconfigurable schedulers accommodating a mixture of real-time, proportional share, fixed priority and other policies, thus overcoming frustrating processor bottlenecks. In this paper we try to overcome this anomaly by proposing an evolutionary strategy for a Meta Hierarchical Scheduler (MHS) in which a user is actively involved in the design cycle of a scheduler. Our framework analyzes user requirements by formulating an abstract model for an optimum scheduler with the help of an evolutionary algorithm that satisfies the needs. Finally a C source code for this MHS is generated using a code generator for the Linux kernel. Our experimental framework demonstrates that our MHS enhances through the evolutionary approach, the user satisfaction level by a factor of two as compared to the satisfaction level achieved by the standard Linux scheduler. Keywords—Adaptable Schedulers, Hierarchical Scheduling, Evolutionary Schedulers, Linux Scheduler, Context-

For a more intensive presentation and an active involvement of students, in particular in large classes, we present a novel approach termed Cooperative Theme and Tool Competence For Learning (CoTTCoL). One of the novelties of our multistep CSCL method is that it utilizes (and advertises) a congruence between objects and operational structures in the theme area and the learning process. To this end combined educational tool and subject area competence are needed. We explain our approach in the example of a new undergraduate class “Operating Systems and Networking”. We report on results, experiences, and project extensions in the works that reach into a large variety of topic areas for teaching.

The major problem for establishing technologies based on solar or wind power, or on renewable energy sources, is an adequate management for the densely networked power sources. This paper describes the first phase of the combined R&D project DEZENT to be funded shortly, between the Schools of Computer Science and Electrical Engineering at the University of Dortmund, devoted to decentralized and adaptive electric power management through a distributed real-time multi-agent architecture. A novel research problem to be specifically faced in this respect is the appropriate handling of unpredictable consumer requests or producer problems. Hence a key result for its solution is that the agents can be most adequately supported through the reactive and adaptable real-time services in the safety-critical operating system MELODY (see [WeL97, WLMR99, WBF01] for selective reading). We briefly outline a two-stage distributed negotiation algorithm, for day-ahead planning and handling of unpredictable power needs and supply situations. Through a 3-year experimental study we will evaluate our claim that both individual/ local consumer and global production and distribution costs will be comparably lower than under a centralized management, let alone their superior flexibility and reliability.

Transaction handling in real-time applications has been studied since quite a number of years, pursuing a variety of research topics. Considering concurrency control, in particular for replicated objects, the algorithms suggested have been moderate modifications of algorithms originally defined for non-real-time database applications. This resulted, for some of them (O2PL and OCC), in a considerable performance improvement (total deadline failure rate) compared to the original versions. As their specific time-sensitive parameter these real-time versions were designed to work on a transaction deadline. While transaction deadlines are to be chosen in accordance with the deadlines of the tasks constituting the transactions no attention has been paid to the latter ones. Taking the perspective of object or task/transaction similarity we experimentally demonstrate that it is highly advantageous (if not indispensable) for the real-time performance to fully utilize the task level and structure of transactions, in particular for safety-critical applications.