Efficient propagation of uncertainties in manufacturing supply chains: Time buckets, L-leap, and multilevel Monte Carlo methods. (2020)

• Record Type: Journal Article
• Authors: Chiang, Nai-Yuan; Lin, Yiqing; Long, Quan
• Is Part Of: Operations research perspectives. Volume 7(2020)
• Publication Date: 2020
• Subjects: Uncertainty modeling -- Discrete event simulation -- Multilevel Monte Carlo -- L-leap -- Supply chain; Operations research -- Periodicals; Management science -- Periodicals; 658.403405
• DOI: 10.1016/j.orp.2020.100144
• Languages: English
• ISSNs: 2214-7160

Abstract

Highlights: We construct a leap method to accelerate the stochastic simulation of manufacturing supply chain in operational research. The method is able to consider production time, transportation time, limited capacity, inventory management, pull system and back orders. We use the MLMC method based on the time buckets to propagate the uncertainties in a supply chain, where most of the computational work is shifted from the expensive models, e.g., DES, to the cheap models defined by large time buckets. The proposed approach is able to match the model accuracy of DES while overcoming its scalability limitation in time with the help of MLMC. Abstract: Uncertainty propagation of large-scale discrete supply chains can be prohibitive when numerous events occur during the simulated period and when discrete-event simulations (DES) are costly. We present a time-bucket method to approximate and accelerate the DES of supply chains. Its stochastic version, which we call the L(logistic)-leap method, can be viewed as an extension of the leap methods (e.g., τ -leap [36]and D -leap [6] developed in the chemical engineering community for the acceleration of stochastic DES of chemical reactions). The L-leap method instantaneously updates the system state vector at discrete time points, and the production rates and policies of a supply chain are assumed to be stationary during each time bucket. We propose using the multilevel Monte Carlo (MLMC) method to efficiently propagate the uncertainties in a supply chain network, where the levels are naturally defined by the sizes of the time buckets of the simulations. We demonstrate the efficiency and accuracy of our methods using four numerical examples derived from a real-world manufacturing material flow application. In these examples, our multilevel L-leap approach can be faster than the standard Monte Carlo (MC) method by one or two orders of magnitude without compromising accuracy.