The EEPR-PAT (Energy-Efficient Purpose and Read-data-from with Prioritization for Aborted Transactions) scheduler is previously proposed in our studies to energy-efficiently serialize transactions by the significancy of purposes while maintaining the read-data-from (RDF) relation of transactions. Furthermore, when a transaction is aborted and restarted, the execution time (ET) of each restarted transaction can be shortened in the EEPR-PAT scheduler. However, the ET of each restarted transaction increases as the number of concurrent transactions increases. In this paper, an EEPR-HPRT (EEPR with High-Prioritization of Restarted Transactions) scheduler is newly proposed to moreover shorten the ET of each restarted transaction. We show the average ET of restarted transactions are moreover shortened in the EEPR-HPRT scheduler than the EEPR-PAT scheduler by a simulation.