In the mobile k-server problem, servers with fixed home locations travel to locations of requests to provide service and then return to their home locations. Given the home locations of the k-servers, the service times, release times and locations of the n requests, we seek to determine a service schedule so that the total waiting time of all requests is minimized. This paper exploits the strong connection between the mobile k-server problem and job scheduling. It presents optimal algorithms for a few special cases of the problem and provides worst-case performance analysis of some heuristics for the general case. For the special cases it is possible to rank order competing sets of home locations with respect to total (or average) waiting time.
