DPO-Tuned Large Language Models for Segmentation in Simultaneous Speech Translation