Description
Laser wakefield acceleration (LWFA) has emerged as a strong candidate for next-generation compact accelerators due to its ability of reaching high accelerating gradients. However, LWFA technique is highly sensitive to the properties of driving laser pulse and small distortions in laser wavefront can significantly affect the wakefield formation and the overall quality of produced electron beam. In this work, we apply Bayesian optimization to adaptively control the laser wavefront using deformable mirrors for achieving high-quality electron beams. Particle-In-Cell simulations are performed using open-source code - WarpX, where Zernike polynomial coefficients serves as optimization variables for parameterizing the laser phase. The optimizer iteratively searches for wavefront configurations to achieve a high-quality beam with enhanced metrics like charge, energy spread and peak energy. The study provides a computational framework for data-driven wavefront optimization in LWFA, with an extension towards experimental implementation planned on our 100 TW laser-facility (UHI100) in the upcoming months.