Framework for improving workflow stability: deployment of optimized capacity buffers in a synchronized construction production

Abstract

Construction sites are dynamic environments due to the influence of variables such as changes in design and processes, unsteady demand, and unavailability of trades. These variables adversely affect productivity and can cause an unstable workflow in the network of trade contractors. Previous research on workflow stability in the construction and manufacturing domains has shown the effectiveness of ‘pull’ production or ‘rate driven’ construction. Pull systems authorize the start of construction when a job is completed and leaves the trade contractor network. However, the problem with pull systems is that completion dates are not explicitly considered and therefore additional mechanisms are required to ensure the due date integrity. On this basis, the aim of this investigation is to improve the coordination between output and demand using optimal-sized capacity buffers. Towards this aim, production data of two Australian construction companies were collected and analyzed. Capacity and cost optimizations were conducted to find the optimum buffer that strikes the balance between late completion costs and lost revenue opportunity. Following this, simulation experiments were designed and run to analyze different ‘what-if’ production scenarios. The findings show that capacity buffers enable builders to ensure a desired service level. Size of the capacity buffer is more sensitive to the level of variability in contractor processes than other production variables. This work contributes to the body-of-knowledge by improving production control in construction and deployment of capacity buffers to achieve a stable workflow. In addition, construction companies can use the easy-to-use framework tested in this study to compute the optimal size for capacity buffers that maximizes profit and prevents late completions.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 26 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.