Optimization approaches for supply network coordination and design project scheduling

Date of Completion

January 2006


Engineering, Electronics and Electrical|Operations Research




The manufacturing industries have gradually shifted from "cost-based competition" and "quality-based competition" to today's "time-based competition" with cost and quality considered given. Accordingly, manufacturers are constantly pressed to have short lead times to rapidly respond to changing customer demands. This caused a major shift on manufacturing strategies, including effective coordination with suppliers for raw materials, component, and parts, and speeding up the product design process by using concurrent engineering. ^ The first part of the study presents an optimization framework for make-to-order supply network coordination. Interactions among organizations are modeled by a set of inter-organization precedence constraints, and the objective is to achieve the organizations' individual and shared goals of fast product delivery and low inventory. These inter-organization constraints are relaxed by using a set of inter-organization prices that represent marginal costs per unit time for the violation of such constraints. The overall problem is thus decomposed into organizational subproblems, where individual organizations schedule their activities based on their internal situations and inter-organization prices. Coordination is achieved through an iterative price updating process carried out in a distributed and asynchronous manner. With prices dynamically updated and schedules adjusted, this approach coordinates activities to fulfill existing commitments while maintaining agility to take on new orders. ^ In the second part of the study, a novel formulation for design project scheduling with explicit modeling of task dependencies and the associated communication under the concurrent engineering paradigm. General dependencies are modeled as combinations of three basic types representing sequential, concurrent, and independent processes. Communication is modeled as tasks, and its interactions with design tasks are described by sets of inter-task constraints to capture the characteristics of concurrent engineering. A solution methodology that combines Lagrangian relaxation, dynamic programming, and heuristics is then developed to schedule design and communication tasks for timely project completion. ^