New method improves traffic model tuning with fewer simulations
FLAT: Smoothing the Rugged Landscape for Learnable, Sample-Efficient Traffic Calibration
Computational Engineering, Finance, and ScienceMachine Learning
Summary
Calibrating detailed traffic simulations is slow and costly because each test requires running a full simulation. The authors developed FLAT, a way to better understand and smooth the relationship between different driving behavior settings and how well the simulation matches real traffic. This makes it possible to learn from fewer tests and focus simulations where they help most. Their method performs better than established techniques across multiple real traffic scenarios.
What this means in practice
- •For traffic simulation developers: Use FLAT to calibrate traffic simulation parameters more accurately with fewer simulation runs in digital twin environments.
- •For urban planning teams: Apply FLAT to quickly tune microscopic traffic models for various real-world scenarios to improve traffic flow predictions.
Authors
Haopeng Deng, Shuo He, Dayuan Wang
Abstract
Calibrating microscopic traffic models for digital twins is an expensive black-box optimization problem: tuning car-following and lane-changing parameters requires a full simulation run, affording only a tight budget per recalibration window. Matching raw trajectories yields a rugged objective that sparse surrogates cannot learn, reducing sequential acquisition to near-random probing. We present FLAT, which couples what to optimize with where to sample next. An eight-dimensional behavioral fingerprint smooths the parameter-error landscape, making the objective learnable from a few dozen samples; annealed lower-confidence-bound (LCB) acquisition then spends each remaining run where it most reduces error. The surrogate, interchangeable among a Gaussian process (GP), random forest (RF), or multi-layer-perceptron (MLP) ensemble, plugs into the same LCB loop. Across six heterogeneous real-world scenes, FLAT-GP achieves the lowest scene-averaged behavioral error, winning 6/6 scenes against SPSA, GA, and CMA-ES and 5/6 against TPE under the matched budget. Some baselines need up to 4.4 times more simulations to match. Ablations show objective choice shifts final behavioral error by 81% on average, removing sequential LCB raises the six-scene mean by 20%, and surrogate choice shifts it by at most 4.2%, confirming gains trace to objective geometry and sequential allocation rather than surrogate capacity.