Michael Schwind - Academia.edu (original) (raw)

Papers by Michael Schwind

This paper presents a decentralized cooperative economic scheduling mechanism for a supply chain ... more This paper presents a decentralized cooperative economic scheduling mechanism for a supply chain environment. For this purpose, we design autonomous agents that minimize the production or transportation and outsourcing costs incurred by the external execution of a task. The decentralized cooperative scheduling approach comprises two parts: the individual optimization of an agent's local schedule and the cooperative contract optimization, either by outsourcing the task or by (re-)contracting the release time and due time with the contract partners aiming to maximize their total profits. A negotiation mechanism based on trust accounts is employed to protect the agents against systematic exploitation by their partners.

Herding combined with threshold behavior is introduced to the Frank- furt Artificial Stock Market... more Herding combined with threshold behavior is introduced to the Frank- furt Artificial Stock Market in the form of a new kind of so called retail agents. Retail agents remain inactive for most of the time and are only activated if an individual threshold of price increases at the exchange is reached. Once activated they imitate profitable strategies of their neighboring agents. Initiated through heavy losses and falling short of cost prices they sell their shares and fall into hibernation and restart their cycle after an individual number of trading days. The relation between the behavior of the retail agents and the stylized facts of the capital markets will be the center of our study.4

Proceedings of the 38th Annual Hawaii International Conference on System Sciences

This paper presents a decentralized negotiation protocol for cooperative economic scheduling in a... more This paper presents a decentralized negotiation protocol for cooperative economic scheduling in a supply chain environment. For this purpose we designed autonomous agents that maximize their profits by optimizing their local schedule and offer side payments to compensate other agents for lost profit or extra expense if cumulative profit is achievable. To further increase their income the agents have to apply a randomized local search heuristic to prevent the negotiation from stopping in locally optimal contracts. We show that the welfare could be increased by using a search strategy similar to Simulated Annealing. Unfortunately, a naïve application of this strategy makes the agents vulnerable to exploitation by untruthful partners. We develop and test a straightforward mechanism based on trust accounts to protect the agents against systematic exploitation. This "Trusted" Simulated Annealing mechanism assures truthful revelation of the individual opportunity cost situation as the basis for the calculation of side payments.

Service-Oriented Computing – ICSOC 2007, 2010

Logistics service providers are forced to optimize their operations due to increasing market pres... more Logistics service providers are forced to optimize their operations due to increasing market pressure. Outsourcing and network optimization are measures to improve their market position. Although the underlying operational research problems in logistics are well documented and the business tendency towards outsourcing already manifests itself in the growth of many local multi vendor logistics providers, the connections between both parts into an innovative and integrated logistics concept is not yet established. Our OPEN Service-Oriented Computing for LOGistics (OPEN SOC LOG) approach proposes a system which integrates state-of-the-art optimization tools with a service-oriented computing concept in order to provide a spectrum of standardized and combinable services for logistics companies ranging from tour and network optimization to contractual and accounting functionalities.

The paper presents a yield optimizing scheduling system (YOSS) in a decentralized supply chain ma... more The paper presents a yield optimizing scheduling system (YOSS) in a decentralized supply chain management (DSCM) environment. For this purpose we employ a DSCM scenario in which each supplier, provides system parts by order of one or more customers. The supplier tries to optimize its return by searching the optimal schedule for the single machine he uses to manufacture the requested products. The schedule is based on the suppliers knowledge about delivery due dates, accomplishment priorities and job prices associated to the production tasks. The optimizer constructs a mandatory schedule by inserting the requested jobs, which arrive stochastically from the customers, gradually into the production queue if the job yields a sufficient return. The tasks are finally executed on the suppliers machine following the fixed queue schedule. The objective of YOSS is to learn an optimal acceptance strategy for the offered jobs. For this purpose YOSS is divided into a deterministic scheduling component (DSC), which does the processing queue assignment of the jobs according to the delivery due date and timeout penalty cost and a reinforcement learning algorithm (RLA), which makes the acceptance decision using a job price-timeout penalty cost difference as optimization criterion. Tier 1

Der Austausch von Ladekapazitäten in Transportbörsen ist ein etabliertes Verfahren der logistisch... more Der Austausch von Ladekapazitäten in Transportbörsen ist ein etabliertes Verfahren der logistischen Praxis. Wenige der meist Web-basierten Marktplätze sind jedoch in der Lage, Synergien zu berücksichtigen, die durch geeignete Kombination von Lieferstrecken verschiedener Transportanbieter entstehen. Eine Methode hierfür ist die Verwendung kombinatorischer Auktionen, bei denen auf Bündel von Streckenabschnitten geboten werden kann. Der Beitrag beschreibt eine kombinatorische Auktion zum innerbetrieblichen Austausch von Logistikdienstleistungen.

The paper presents a decentralized supply chain management approach based on reinforcement learni... more The paper presents a decentralized supply chain management approach based on reinforcement learning. Our supply chain scenario consists of loosely coupled yield optimizing scheduling agents trying to learn an optimal acceptance strategy for the offered jobs. The optimizer constructs a mandatory schedule by inserting the requested jobs, which arrive stochastically from the customers, gradually into a production queue if the job yields a sufficient return. To reduce complexity the agents are divided into three components. A supply chain interface, classifying job offers, a reinforcement learning algorithm component, which makes the acceptance decision and a deterministic scheduling component, which processes the jobs and generates a preliminary state space compression. The reinforcement learning algorithm accepts offers according to their delivery due date, the job price, the timeout penalty cost, and the information provided by the scheduling component. The tasks are finally executed on the suppliers machine following the queue's schedule. In a performance comparison of our yield optimizing agent it turns out that the reinforcement learning solution outperforms the simple acceptance heuristic for all training states. production and transportation costs using currency exchange rates, tariffs, production-, inventory, late delivery and transportation costs, the RLA chooses between three possible suppliers and one of two transportation modes. SMART is tested by applying various demand patterns to the supply chain, grounding on a Erlang probability distribution modified by diverse mean and deviation parameters. Compared with two heuristics, one called LH standing for local heuristic preferring a inner country production and distribution policy and another denoted as BH balanced heuristic while issuing demand mainly to warehouses with low capacity load, the SMART allocation mechanism provides the highest reward. Tier x+1 Agentx+1,

Inspired from the research of Frank Westerhoff from the University of Osnabrück about heterogeneo... more Inspired from the research of Frank Westerhoff from the University of Osnabrück about heterogeneous traders and the Tobin tax and the expertise from Paul Spahn, University of Frankfurt, by order of the German government, about the feasibility of an exchange rate transac- tion tax, the authors wanted to support the results of Frank Westerhoff within a different model setting and wanted to extend his approach further by incorporating the approach of Paul Spahn's Tobin-cum-circuit-breaker-tax. The results point into the same direction as Westerhoff's simula- tions. The Tobin-cum-circuit-breaker-tax of Paul Spahn showed a superior performance over the standard Tobin tax.

… European Conference on …, 2002

The focus of this paper is the design of a mechanism to help economic agents-either autonomously ... more The focus of this paper is the design of a mechanism to help economic agents-either autonomously or cooperatively planning-to achieve Pareto-optimal allocation of resources via a completely decentralized coordination of a logistics network. Besides giving a classification and a short review of existing scheduling approaches capable for supply chain management, this article specifies and evaluates protocols employing time-depended price-functions. By performing simulations with the implemented protocol using well-known Job Shop Scheduling Problems as a benchmark we show the efficiency and feasibility of the designed mechanism. The approach enables each agent to exploit the external effects caused by resource constraints of its supply chain contractors by adapting its production planning. Additionally the systems capability to reconfigure itself in case of production resources failure is increased. The evaluation of the protocols concludes with a welfare analysis investigating the payoff distribution along the supply chain. Finally we conclude that future research on this topic should turn to learning agent systems to reduce communication costs.

International Journal of Production Economics, 2012

We present a location routing model that considers the requirements of real-world small package s... more We present a location routing model that considers the requirements of real-world small package shippers. It integrates the choice between self-operating and subcontracting a depot. Furthermore, we consider the relocation costs which occur when established networks are restructured. We present a hybrid heuristic solution method based on simulated annealing for the location phase and variable neighborhood search for the routing. Through extensive numerical studies we show the competitiveness of our solution method on standard benchmarks and analyze the influence of subcontractors on the network design of a small package shipper.

Electronic Markets, 2008

His research focuses on dynamic pricing methods and market engineering, especially combinatorial ... more His research focuses on dynamic pricing methods and market engineering, especially combinatorial auctions for resource allocation and procurement processes.

WIRTSCHAFTSINFORMATIK, 2006

The study of network topologies provides interesting insights into the way the principles on whic... more The study of network topologies provides interesting insights into the way the principles on which the construction of connected systems are based influence diffusion dynamics and communication processes in many socio-technical systems. & Empirical research has shown that there are principles for the construction of social networks and their technical derivatives, like e-mail networks, the Internet, publication coauthoring, or business collaboration. & Such real world networks attach new members over time and the mode of attachment prefers existing members that are already well connected. This principle is called "preferential attachment" and leads to the emergence of "scale-free" networks. & Scale-free networks seem to be a better fit for the description of real world networks than the random networks used so far. Their behavior in terms of diffusion and communication processes is fundamentally different from that of random networks. & To illustrate the value of scale-free networks for applications in information systems research, examples will be given to illustrate their usefulness for real world network modeling. A communication network of security traders will show what impact network topology has on the dynamics of complex socio-technical systems.

The 2003 Congress on Evolutionary Computation, 2003. CEC '03.

This paper presents and compares three heuristics for the combinatorial auction problem. Besides ... more This paper presents and compares three heuristics for the combinatorial auction problem. Besides a simple greedy (SG) mechanism, two metaheuristics, a simulated annealing (SA), and a genetic algorithm (GA) approach are developed which use the combinatorial auction process to find an allocation with maximal revenue for the auctioneer. The performance of these three heuristics is evaluated in the context of a price controlled resource allocation process designed for the control and provision of distributed information services. Comparing the SG and SA method shows that depending on the problem structure the performance of the SA is up to ¢ ¡ ¤ £ higher than the performance of the simple greedy allocation method. The proposed GA approach, using a random key encoding, results in a further improvement of the solution quality. Although the metaheuristic approaches result in higher search performance, the computational effort in terms of used CPU time is higher in comparison to the simple greedy mechanism. However, the absolute overall computation time is low enough to enable real-time execution in the considered IS application domain. P. If not all requested resources are assigned to him, the partial acquisition of the resources has much less value because of production delays or even production failures. This results in the higher than linear (superadditive) valuation of the bundled goods compared to the single items valuation.

2011 44th Hawaii International Conference on System Sciences

Electronic transportation marketplaces (ETMs) are inter-organizational information systems that p... more Electronic transportation marketplaces (ETMs) are inter-organizational information systems that provide an eco- nomic platform for the exchange of logistics services between shippers, carriers, and freight forwarders. Despite the fact that some of these marketplaces have already been in operation for over twenty years they have not achieved any significant influence in the logistics services domain until now. In a survey

Proceedings of the Eighth International Conference on Enterprise Information Systems, 2006

Our paper presents an agent-based simulation environment for task scheduling in a distributed com... more Our paper presents an agent-based simulation environment for task scheduling in a distributed computer systems (grid). The scheduler enables the simultaneous allocation of resources like CPU time, communication bandwidth, volatile and non-volatile memory while employing a combinatorial resource allocation mechanism. The resource allocation is performed by an iterative combinatorial auction in which proxy-bidding agents try to acquire their desired resource allocation profiles with respect to limited monetary budget endowments. To achieve an efficient bidding process, the auctioneer provides resource price information to the bidding agents. The calculation of the resource prices in a combinatorial auction is not trivial, especially if the the bid bundles exhibit complementarities or substitutionalities. We propose an approximate pricing mechanism using shadow prices from a linear programming formulation for this purpose. The efficiency of the shadow price-based allocation mechanism is tested in the context of a closed loop grid system in which the agents can use monetary units rewarded for the resources they provide to the system for the acquisition of complementary capacity. Two types of proxy-bidding agents are compared in terms of efficiency (received units of resources, time until bid acceptance) within this scenario: An aggressive bidding agent with intense rising bids and a smooth bidding agent with slow increasing bids. tomized stock chart data or the broadcast of a public event to viewers via TCP/IP protocol.

Advances in Production, Logistics and Traffic, 2019

Under the light of recent studies, it seems that not only the structure and topology of networks ... more Under the light of recent studies, it seems that not only the structure and topology of networks but also their growth follows some laws of nature. It appears that this statement is also true for transport and logistics networks. Moreover, specific attributes of a given network like small-world and scale-free properties have additional influences on the efficiency of a network. The principal concern of this study is to assess the topology and structure of the different types of weighted and unweighted networks of “Deutsche Bahn (DB)” or German Railway and to evaluate their local and global efficiencies. Additionally, the different types of efficiencies of a network are used to develop an algorithm that will further help in optimising the network in the case of some planned addition or alteration of edges to the underlying or basic topology of the network. The assessment processes are divided into two parts and carried out on three different networks (both weighted and unweighted) na...

36th Annual Hawaii International Conference on System Sciences, 2003. Proceedings of the, 2003

The economic impact of the growth dynamic of standards is often described from a macroeconomic po... more The economic impact of the growth dynamic of standards is often described from a macroeconomic point of view, employing network effect theory and models dealing with externalities. Game-theoretic models try to depict and predict the situation on the microeconomic side. We follow a new approach, which simulates the system's behavior based on the modeling of a set of individual conduction rules, and their interaction in a closed environment. Implementing such an Agent based Computational Economics approach, using a simulation environment called SWARM, we assume the existence of three firm sizes combined with three types of standards. Each standard has an optimal fit to a firm size, which results in reduced costs according to standardization benefits, whereas other combinations lead to lower savings respectively. In addition we postulate initial standardization cost for internal restructuring measures and different scopes of communication fitted to the three firm types. Each firm can repeatedly decide to standardize in various simulation passes depending on an expected standardization benefit. As an outcome of various simulations, we observe a dominance of the communication standard preferred by the large companies as a function of growing network density. For an increasing communication range the same behavior emerges for all firm types. The model underpins the well known concentration tendency in real worlds' technology markets.

2011 IEEE 13th Conference on Commerce and Enterprise Computing, 2011

ABSTRACT

2012 45th Hawaii International Conference on System Sciences, 2012

Despite the fact that the vehicle routing problem (VRP) with its variants has been widely explore... more Despite the fact that the vehicle routing problem (VRP) with its variants has been widely explored in operations research, there is very little published research on the VRP concerning real world constraint combinations and large problem sizes. In this work a heuristic solution approach for the VRP with real world constraints is presented driven by the requirements defined by clients in the courier, express and parcel (CEP) delivery industry in order to support their routing plan decisions and driver assignments. The solution algorithm used combines several local-search-based heuristics with constructive elements to solve the VRP with driver learning (VRPDL). As conceptual proof large instances for the capacitated VRP (CVRP) including 560 to 1200 customers are tested and compared to known benchmark results. From those instances new sub-instances are created and sequentially tested adding the driver learning constraint. Finally, the solver is applied to real world CEP instances with driver learning.

Lecture Notes in Computer Science, 2006

This paper presents an agent-based simulation environment for task scheduling in a grid. Resource... more This paper presents an agent-based simulation environment for task scheduling in a grid. Resource allocation is performed by an iterative combinatorial auction in which proxy-bidding agents try to acquire their desired resource allocation profiles. To achieve an efficient bidding process, the auctioneer provides the bidding agents with approximated shadow prices from a linear programming formulation. The objective of this paper is to identify optimal bidding strategies in multi-agent settings with respect to varying preferences in terms of resource quantity and waiting time until bid acceptance. On the basis of a utility function we characterize two types of agents: a quantity maximizing agent with a low preference for fast bid acceptance and an impatient bidding agent with a high valuation of fast allocation of the requested resources. Bidding strategies with varying initial bid pricing and different price increments are evaluated. Quantity maximizing agents should submit initial bids with low and slowly increasing prices, whereas impatient agents should start slightly below market prices and avoid ‘overbidding’.