Improvement of RMC delivery operation performance through inventory control based on discrete event simulation

Abstract

As a highly demand construction material, ready-mixed concrete (RMC) is a common commodity in the construction industry. Because of its inherent characteristics and strict quality requirements, the production, delivery and placement process of RMC is a topic of interests to both contractors and material suppliers. Among the factors affecting the RMC supply performance, the dispatching schedule of delivery trucks has a direct effect on the total operation time and queuing time. The primary obstacles in determining optimal dispatching intervals are uncertainties in traffic and unloading activity. To offset the negative impacts of these uncertainties, it is important to set an inventory of delivery trucks in the supply chain is necessary. However, although inventory can ensure continuity, it can also lead to waste to a certain extent. Therefore, finding a cost-effective inventory control mode is necessary to improve the performance of the supply chain. This thesis develops a model where inventory control policies working on an RMC delivery process with the aim of optimizing the supply chain from the perspectives of both the contractor and supplier. This model will: (1) maintain the continuous work, (2) reduce the wasted time at the construction site, and (3) allocate trucks more efficiently. Data collection and implementation are based on a major building project in United Arab Emirates which required large amount of RMC. And the observed process consists of 502 RMC delivery cycles. On the basis of original model, two control policies are implemented in the model and compared in terms of their impact on process behaviour. The discrete event simulation (DES) model results illustrate the advantages of flexible dispatching intervals provided by fixed-quantity inventory control policy including maintaining continuous work and reducing waste. An alternative solution to the queuing problem is also discussed by comparing queuing model outputs to discrete event simulation outputs. The results show that the queuing model fails to provide an accurate estimation of queuing statues when managerial actions are involved. The decision makers, plant managers, and contractors could potentially take advantage of the findings of this thesis to better understand and manage their supply chain.