Operations Research/Computer Science Interfaces Series

Transcription

1 Operations Research/Computer Science Interfaces Series Volume 50 Series Editors: Ramesh Sharda Oklahoma State University, Stillwater, Oklahoma, USA Stefan Voß University of Hamburg, Hamburg, Germany For further volumes:

2

3 Jörn Schönberger Model-Based Control of Logistics Processes in Volatile Environments Decision Support for Operations Planning in Supply Consortia 123

4 Jörn Schönberger ChairofLogistics University of Bremen Wilhelm-Herbst-Str Bremen Germany jsb@uni-bremen.de ISSN X ISBN e-isbn DOI / Springer New York Dordrecht Heidelberg London Library of Congress Control Number: c Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (

5 Preface The overarching goal of this study is to report on the development of new and innovative ideas to support the integration of decision making between a coordinator in a supply consortium (the principal) and a subordinate service providing partner (the subordinate agent). In particular, new approaches for the management of logistics processes in volatile and uncertain planning environments are developed and evaluated. These approaches improve the common decision making between a supply consortium coordinator and service proving partner(s), which finally leads to an increase of the quality of the generated value creation process provided by the supply consortium. An installation and a setup of opportunities for the coordinator to intervene into the local resource deployment of a service providing partner is the core idea behind these new approaches. If necessary from the viewpoint of the complete supply consortium, the coordinator should be allowed (and obliged) to control temporarily the dispatching of resources partly or completely. The motivation behind this intervention is to protect and stabilize the performance of the overall supply consortium if a certain partner does not act in the sense of the consortium. To realize and to implement the coordinator interventions, we propose that the superior coordinator manipulates the formal deployment decision model of the subordinate service providing agent. Critical changes of the decision situation (demand peaks, resource unavailability,...) are reflected back into the formalized representation (decision model) of the subordinate agent s decision task. Answers to the following research questions are sought in this study. 1. What are the state-of-the-art techniques to integrate decision making of principals and service agents in a supply consortium? 2. In which situations can these techniques be applied successfully and in which situations do they fail? 3. Why do the so far discussed decision supporting approaches fail? 4. How can their deficiencies be remedied? 5. How can the improved decision support approaches contribute to handling more dynamically appearing disturbances? v

6 vi Preface 6. Can the stability of once made decisions be increased? 7. How high is the price for the extension of the abilities of the integration techniques? We have defined three milestones, each labeling a significant progress in the reported study. The first milestone is the identification of limits and barriers of repeated decision making (online decision making) in volatile and uncertain environments in presence of principal-agent-relationships. By means of an exemplarily analyzed interaction of a supply consortium coordinator (principal) and a transport service agent, the conceptual limitations of the use of model-based decision support for principal-agent interactions are explored (Part I). In Chapter 2, the determination problems and the subsequent update problems of freight transport processes in volatile and uncertain planning environments are addressed. The ability to cope with spontaneous and unexpectedly appearing demand peaks is identified as a core challenge. We introduce a corresponding version of the online vehicle routing problem with time windows in order to prepare a process management simulation. Potentially conflicting planning objectives of the supply consortium coordinator and of the transport organizing fleet manager are identified and analyzed. State-of-the-art concepts for integrating the planning objective and decision making of a superior coordinator and a transport fleet managing agent are presented in Chapter 3. Within computational experiments these concepts are evaluated and their limitations are revealed. The second milestone comprises the deficiency analysis of recent integrated decision making techniques and the derivation of a conceptual framework for remedying the identified shortcomings (Part II). Chapter 4 addresses the extension of adaptive process control systems by installing coordinator intervention features. We propose a close-control-circuit to alter the agent s formal process decision model, e.g., a mathematical optimization model. The necessary modifications of the objective function and of the constraint set are feedback-driven. The feedback of a process is determined by comparing the current process quality and a given reference process quality. In Chapter 5, we derive three model-adjusting strategies for the aforementioned online vehicle routing problem with time windows. Within comprehensive computational simulation experiments, we identify the best possible parameterization of the strategies. Furthermore, we compare these three strategies among themselves and in addition with the state-of-the-art techniques describes in the third chapter. It turns out that the proposed techniques are able to remedy the deficiencies of the so far known techniques. The third and final milestone consists in the analysis of the impacts of applying the new techniques to dynamic transport process planning problems in volatile environments (Part III). Flexibility of transport systems controlled by the integrated strategies is investigated in Chapter 6. We interpret flexibility as the property to handle unexpected disturbances of processes so that an update of existing processes is possible. In Chapter 7, we try to protect once made decisions from further revisions in order to reduce the process nervousness. It is shown that the application of an integrated decision making between a coordinator and a transport service providing agent contributes to prevent further decision revisions, so that once accepted

7 Preface vii decisions exhibit an increased stability. Finally, we investigate the tradeoff between process quality increase (as a result of the application of the integrated planning approaches) and the additional expenditures resulting from overruling the typically cost-minimal deployment decisions of a service agent (Chapter 8). We conclude this study with a summary of the main findings (Chapter 9). This book reports scientific results from the research project Autonomous Adaptation of Vehicle Schedules. The reported research was carried out by the group of the Chair of Logistics, University of Bremen. Prof. Dr.-Ing. Herbert Kopfer, holder of the Chair of Logistics and principal investigator in this project, gave me the opportunity to design and to conduct my project-related research as freely as possible. He supported me continuously during the six years of project work. The present research stands to benefit from continuous scientific discussions. In this context I would like to express my gratitude to former and current colleagues at the Chair of Logistics, to the associates from the Collaborative Research Center 637 at the University of Bremen and to the colleagues who provided my with new ideas when we met at conferences and workshops. In particular, I have to thank Prof. Dr. Hans-Dietrich Haasis (Institute of Shipping Economics and Logistics), Prof. Dr. Christian Bierwirth (Martin-Luther-Universität Halle-Wittenberg), Prof. Dr. Martin G. Möhrle (University of Bremen) and Prof. Dr. Thorsten Poddig (University of Bremen). This research was supported by the German Research Foundation (DFG) as part of the Collaborative Research Center 637 Autonomous Cooperating Logistic Processes (Subproject B7). Bremen, February 2011 Jörn Schönberger

8

9 Contents 1 Process Planning in Supply Consortia Value Creation in Supply Consortia Supply Consortia Resource Deployment Paradigms Centralized Planning Autonomous Control Hierarchical Planning for Consortium-Type Governance Transport Processes in Supply Consortia: Relevance and Challenges Transport Process Derivation in a Supply Consortium Process Endangering Events and Data Uncertainty Part I Model-Based Transport Process Planning: Approaching the Limits 2 Transport Processes and Uncertainty Formalization of Uncertainty within Decision Problems Operational Management of Peaks in Transport Demand A Dynamic Vehicle Routing Problem with Subcontractor Options General Scenario Outline Coordinator s and Service Agent s Interest Integration Reference Configuration Penalization of Late Arrivals Dispatching Task of the Fleet-Managing Agent Generation of Parameterized Test Cases Conclusions Decision Support: Applying the State-of-the-Art Decision Support Systems and Transport Process Planning Process Control by Decision Support Systems DSS-based Control of Transport Process Event-Handling in DSS for Dynamic Transport Dispatching ix

10 x Contents Challenges in Dynamic Disposition and Dispatching Rolling Horizon Disposition and Dispatching Online Planning in HARD- and PEN-Configurations Memetic Algorithm for the Transportation Plan Generation Constraint Handling Techniques Simulation Experiments Report Simulated Scenarios Performance Indicators Process Quality Indicators Financial Process Evaluation Indicators Computational Assessment of HARD and PEN Online-Process Quality Assessment Offline Process Quality Assessment Online-Evaluation of the Process Costs Offline-Assessment of the Process Costs Varying Penalties Conclusions Part II Extending the Application Boundaries of Model-Based Planning 4 Decision Support in Principal-Agent-Relationships Adaptive Process Control Systems Reliable and Unreliable Adaptive Systems Process Controller of Adaptive Systems System Intervention by Image Modification Image Modification and Event Handling in DSS Image Modification and Process Re-Planning Image Modification and Integrated Principal-Agent Resource Allocation Preparing Integrated Principal-Agent Resource Allocation Algorithmic Update of the Agent s Decision Model Summary Adaptive Controllers for Mathematical Optimization Models Preparations System Development Corridor Definition of a Suitable Intensity Function Accounting Scheme Adaptation Accounting Schemes Objective Function Re-Parameterization Adaption of the Accounting Scheme to the Current Process Punctuality Adaptive Exercising of LSP-Options Decision Model Preprocessing and Presolving... 88

11 Contents xi An Adjustable Constraint Family Preparing the Coordinator Agent s Interventions Intervention Specification Random Request Sequencing (RRS) Distance-to-be-Bridged Sequencing (DBS) Vehicle Availability Sequencing (VAS) Remaining Time Based Sequencing (RTS) Isolation Based Sequencing (IBS) Evaluation and Assessment of the Model Controllers Simulated Scenarios Presentation and Discussion of Results Parameterization of SDAD and CSAD Online Comparison of Static and Adaptive Strategies Offline Comparison of Static and Adaptive Strategies Online-Evaluation of the Process Costs Offline Cost Comparison of the Four Strategies Hybridization of Model Adaptation Strategies Parameterization of the Hybrid Strategy Online Performance Comparison Offline Assessment Cost Evaluation Summary of Findings Part III Adaptive Model Controllers in Action 6 Responsiveness Improvement Flexibility and Logistic Operations Literature Review Flexible Plans and Flexible Systems Planflexibility Systemflexibility Quantification of Flexibility Measures for Planflexibility Systemflexibility Quantification Computational Experiments Experimental Setup Results Conclusion of Findings Nervousness Reduction in Re-Scheduling External and Internal Nervousness Schedule Transition Nervousness Transport System Nervousness...126

12 xii Contents 7.4 Flexibility and Nervousness of Logistic Operations Numerical Experiments Conclusions Impacts on Robustness Robustness in the Literature Robustness of Schedules Robustness of Systems Achieving, Implementing and Conserving Robustness Measuring and Quantification of Robustness Planrobustness Systemrobustness Definition of Robustness Basic Terms Evaluation Schemes and Acceptable Updates Evaluation of Disturbances Controlling the Variation between the Original and its Update Referential Variation Comparison of Input-Output-Variations Comparison of Update and Original Comparison of Updates with a Fixed Reference Schedule The Role of the Schedule Update Strategy Robustness of Schedules and Systems Robust Schedules Robust Systems Robustness and Flexibility Quantification of Robustness Planrobustness Quantification Quantification of Systemrobustness Robustness in a Transportation Scenario Setup of the Simulation Experiments Presentation and Discussion of Simulation Results Conclusions Summary and Conclusions Principals, Agents and Dynamic Decision Problems Innovative Methods for Decision Derivation and Evaluation Principal-Agent Relationships: Improved Process Quality References Index...177