1: \begin{abstract}

2: 	The \srs problem is a two-stage problem where, given $m$ machines, jobs must be grouped into

3: 	at most~$m$ bags while the processing speeds of the

4: 	machines

5: 	are unknown. After the speeds are revealed, the grouped jobs must be assigned

6: 	to the machines without being separated. To evaluate the performance of

7: 	algorithms, we determine upper bounds on the worst-case ratio of the

8: 	algorithm's makespan and the optimal makespan given full information. We refer

9: 	to this ratio as the robustness factor. We give an algorithm with a robustness

10: 	factor~$2-1/m$ for the most general setting and improve this to~$1.8$ for

11: 	equal-size jobs. For the special case of infinitesimal jobs, we give an

12: 	algorithm with an optimal robustness factor equal to~\mbox{$e/(e-1)\approx 1.58$}. The

13: 	particular machine environment in which all machines have either speed~$0$

14: 	or~$1$ was studied before by Stein and Zhong~(SODA 2019). For this setting, we

15: 	provide an algorithm for scheduling infinitesimal jobs with an optimal robustness

16: 	factor of~$(1+\sqrt{2})/2\approx 1.207$. It lays the foundation for an

17: 	algorithm matching the lower bound of~$4/3$ for equal-size~jobs. 

18: \end{abstract}