Abstract
In the recent literature, many definitions of partial randomness of reals have been proposed and studied rather discretely. For instance, it is known that for a computable real \epsilon\in(0,1), strong Martin-L\ddot{o}f \epsilon-randomness is strictly stronger than Solovay \epsilon-randomness which is strictly stronger than weak Martin-L\ddot{o}f \epsilon-randomness. In the present work, we firstly give several new definitions of partial randomness --- strong Kolmogorov \epsilon-randomness and weak/strong DH-Chaitin \epsilon-randomness. Then, we investigate the relation between \epsilon-randomness by one definition and \epsilon'-randomness by another. Finally, we show that all of the known definitions of \epsilon-randomness are quasi-equivalent.