Stochastic simulation under input uncertainty: A Review

Abstract

Stochastic simulation is an invaluable tool for operations-research practitioners for the performance evaluation of systems with random behavior and mathematically intractable performance measures. An important step in the development of a simulation model is input modeling, which is the selection of appropriate probability models that characterize the stochastic behavior of the system inputs. For example, in a queueing-system simulation, input modeling includes choosing the probability distributions for stochastic interarrival and service times. The lack of knowledge about the true input models is an important practical challenge. The impact of the lack of information about the true input model on the simulation output is referred to as ‘input uncertainty’ in the simulation literature. Ignoring input uncertainty often leads to poor estimates of the system performance, especially when there is limited amount of historical data to make inference on the input models. Therefore, it is critically important to assess the impact of input uncertainty on the estimated performance measures in a statistically valid and computationally efficient way. The goal of this paper is to present input uncertainty research in stochastic simulations by providing a classification of major research streams and focusing on the new developments in recent years. We also review application papers that investigate the value of representing input uncertainty in the simulation of real-world stochastic systems in various industries. We provide a self-contained presentation of the major research streams with a special attention on the new developments in the last couple of years.

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 — 58 citations exist upstream.

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