9/11/2001,Fall 2001, Copy Right P. B. Luh,1,Modeling and Optimization of Complex Networked Systems: Applications to Operations Scheduling and Supply Network Management,Peter B. Luh, Visiting Professor Center for Intelligent and Networked Systems Department of Automation Room 503, Main Building Tsinghua University Phone: 6279-2438 Email: Peter.Luhuconn.edu,9/11/2001,Fall 2001, Copy Right P. B. Luh,2,Are you in the right classroom? What are complex networked systems?,Computer and communication networks Power systems Supply chains Why are we interested in complex networked systems?,As a person, a team, a department, or a company Mediocrity or excellence? Surviving or thriving? Need to accurately describe the system under consideration and prudent decision-making to maximize an objective(s),Major areas of development and advancement Why do we need to study their modeling and optimization?,9/11/2001,Fall 2001, Copy Right P. B. Luh,3,The major segments in the course Mathematical optimization concepts and algorithms, focusing on methods that can solve practical large-scale problems (7 lectures) Operations scheduling for manufacturing and service sectors, making use of optimization methods learned in the first segment (4 lectures) Strategy, planning, and operations of modern supply networks, exploring distributed and collaborative decision-making and optimization within a networked environment (4 lectures) Participants who are interested in research can take part in extra discussion sessions on mini research projects,9/11/2001,Fall 2001, Copy Right P. B. Luh,4,What do you expect to learn from this course?,Solid understanding of the above subjects Classroom participation experience Mini (or major) research opportunities Improved English What is the workload?,Slightly above average, and intensive in September (8 lectures) and October (7 lectures) (and in November and in December) Prerequisites: A first year graduate course on system theory How can we make it a fun course for all of us?,Preparation: Before and after each class Participation: In class,9/11/2001,Fall 2001, Copy Right P. B. Luh,5,Classes In September: Monday 8:00 9:35, Thursday 8:00 9:35, and Friday 13:30 15:05 In October: Monday 8:00 9:35 and Friday 13:30 15:05 Office Hours: Monday 16:00 17:00, Thursday 16:00 17:00 Time for voluntary discussion sessions to be arranged Major references Dimitri P. Bertsekas, Nonlinear Programming, Second Edition, Athena Scientific, Belmont, MA, 1999 Michael Pinedo and Xiuli Chao, Operations Scheduling with Applications in Manufacturing and Services, McGraw-Hill, 1999 Sunil Chopra and Peter Meindl, Supply Chain Management: Strategy, Planning and Operations, Prentice Hall, 2000,9/11/2001,Fall 2001, Copy Right P. B. Luh,6,Tentative Outline 1. Introduction and Unconstrained Optimization (2.33 lectures) 2. Optimization over a Convex Set (2 lectures) 3. Duality and Dual Methods (2.66 lectures) 4. Operations Scheduling with Applications in Manufacturing and Services (1.33 lectures) 5. Job Shop Scheduling (1.33 lectures) 6. Scheduling Models in Service Industries (1.33 lectures) 7. Supply Chain Management: Strategy, Planning, and Operations (2 lectures) 8. Distributed and Collaborative Decision-Making in Supply Networks (2 lectures),9/11/2001,Fall 2001, Copy Right P. B. Luh,7,Grading: Homework 45% Mid Term 25% Term Project 25% Classroom Participation 5% Total 100% General Rules: Term projects can be done individually or in teams of two on relevant optimization topics or applications. Topics should be based on at least two recent papers. Numerical implementation and testing are strongly encouraged. Term project proposals are due on Friday October 12. The date of presentations and the date of submitting the final reports will be determined later.,9/11/2001,Fall 2001, Copy Right P. B. Luh,8,Homework problems should be clear, concise, and complete. Not all the homework problems will be graded. Grading will be based on some randomly selected problems. Late assignments will be discounted 10% a day, up to 5 days. This policy will be strictly enforced. Homework solutions will be provided on the web sites a week after the due date. Comments and discussions are encouraged in class, after class, during office hours or by appointment. NO CHEATING!,9/11/2001,Fall 2001, Copy Right P. B. Luh,9,Reading Assignment: Bertsekas Sections 1.1, 1.2, and Appendices A and B Today: Introduction and Unconstrained Optimization Motivation and Course Overview Problem Classification Optimality Conditions for Unconstrained Optimization Gradient Methods Framework Tomorrow: Bertsekas Sections 2.3 2.6,9/11/2001,Fall 2001, Copy Right P. B. Luh,10,Mathematical Optimization Concepts and Algorithms 1.2 Problem Classification,General Formulation Minimize f(x), subject to x X Rn f(x): Cost function x: Decision variable X: Constraint set Example 1 A small shop specializes in making 2 types of auto parts,9/11/2001,Fall 2001, Copy Right