Dynamic Control of Flexible Queueing Networks with Application to Shipbuilding.

Abstract

Shipbuilding is a unique industry that uses a wide variety of manufactured components and requires a large number of workers possessing various skills as well as specialized facilities. In recent decades, the U.S. Naval shipbuilding industry has faced challenges of building ships on-time and within budgeted cost. The main problems in Naval ship production are high variability in production workload, ineffective production control, and low facility utilization. Our research emphasizes an innovative production system design and control to improve the shipbuilding production performance. We introduce (1) operational flexibility at the execution level and (2) the release policy Constant Work in Process (CONWIP) concepts at strategic level to shipbuilding. The systems are formulated as flexible queueing networks, and Markov Decision Process (MDP) approach is applied to gain structural insights into the optimal control policy and to optimize the key performance measures such as cost, throughput, and cycle time. Results show that the flexibility bring the robustness to the system which mitigates the variability of the block processing time and also achieve a quicker response to the workload change. We also develop efficient control policies to control the flexible resource which minimize the average holding cost of ship blocks and also improve the system throughput. Another research area we investigate is the outfitting process in shipbuilding. The outfitting refers to the process of fabrication and installation of non-structural components, and represents as much as 50% of the cost of the ship, and up to 50% of ship construction time in many instances. However, due to disturbances from unexpected delays, system variations, capacity limitations, and technological constraints, scheduling of outfitting processes is very complex and can delay the entire ship production system. Therefore, a strategic level planning model and a dynamic control model are developed to provide analytical information which improves the planning and controlling of the outfitting process.