Reference dataset of multi-objective and multi-fidelity optimization in laser-plasma acceleration

2443 particle‑in‑cell simulations used in Bayesian optimization

datasets submitted

This dataset comprises 2443 FBPIC particle‑in‑cell simulations of a laser wakefield accelerator, selected using a Bayesian optimizer. Inputs (n, 5): [plasma density, upramp length, laser focus, downramp length, fidelity]. Objectives (n, 3): [total charge, distance from target energy, bandwidth]. Includes multiple fidelities for surrogate modeling and multi‑fidelity optimization.