Bayesian Precalibration of a Large Stochastic Microsimulation Model
DOI: 10.1109/tits.2014.2304394
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the computational and labor-intensive challenges of calibrating large stochastic traffic microsimulation models. Traditional calibration methods, often referred to as "expert tuning," rely on iterative, manual adjustments of numerous interrelated parameters, a process that is inefficient and struggles to manage complex parameter interactions. To overcome these limitations, the authors propose a fast, iterative probabilistic pre-calibration framework known as "history matching." This approach aims to rule out implausible regions of the parameter space by using emulators—statistical surrogate models—that approximate the simulator’s behavior, thereby significantly reducing the number of required simulator runs while accounting for uncertainty. The methodology involves an iterative scheme comprising expert elicitation of inputs and uncertainties, experimental design, simulator evaluation, emulator fitting, and implausibility calculation. The authors apply this framework to a real-world S-Paramics microsimulation model of a 58km section of the M40 motorway in the UK, which models over 50,000 trips. The process was executed in three waves. Wave 1 utilized a maximin Latin Hypercube design with 484 points to establish initial Gaussian Process emulators. Wave 2 refined the search by evaluating implausibility across the parameter space, identifying 233 non-implausible points for further simulator evaluation with increased replication to better estimate variance. Wave 3 performed final refinements. The emulators modeled the relationship between 25 critical input parameters (including junction calibration, route behavior, and vehicle dynamics) and turn count observations at nine locations across four time steps. The results demonstrate that the pre-calibration framework successfully reduced the volume of the plausible parameter space to 3% of the original domain. Visualization of the implausibility space revealed specific parameter constraints, such as headway and motorway cost thresholds, and highlighted interactions between parameters that univariate analysis missed. Crucially, the pre-calibrated model improved the fit to observational data compared to the traditional expert-tuned model. For instance, while the expert-calibrated model consistently underestimated turn counts during peak hours, the pre-calibrated runs achieved significantly smaller errors. The study also noted that for some locations, the simulator could not match observations under any tested parameters, indicating potential structural errors in the model itself. The significance of this work lies in demonstrating that automatic, simultaneous calibration of all parameters in a complex stochastic simulator is feasible and superior to manual tuning. By using emulators, the method drastically cuts computational costs, allowing for the exploration of parameter interactions and uncertainty quantification that are impractical with direct simulation. This approach provides a robust foundation for subsequent formal Bayesian calibration and offers transport planners a more efficient, evidence-based tool for validating and improving traffic simulation models.
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | OpenAlex-citations | — | — | 1 | 2026-06-19 |
| archive | success | unpaywall | — | — | 2 | 2026-06-25 |
| extract | success | cached | — | — | 2 | 2026-06-26 |
| clean | success | clean | — | — | 1 | 2026-06-19 |
| chunk | success | chunk | — | — | 1 | 2026-06-19 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-19 |
| promote | success | — | — | — | 1 | 2026-06-19 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-26 |
| tag | success | vector_similarity | — | — | 6 | 2026-06-19 |
| verify | success | — | — | — | 1 | 2026-06-26 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Theoretical Contribution: computational model