Imagine that there exists an algorithm SPLITk that can split a list L of n elements into k sub lists, each containing one or more elements, such that sub list i contains only elements whose values are less than all elements in sub list j for i

(a) What is the worst-case asymptotic running time for SORTk? Why?
(b) What is the average-case asymptotic running time of SORTk? Why?