A Queueing Network Based Approach to Distributed Aircraft Carrier Deck Scheduling
Rajarshi Ghosh Dastidar, Emilio Frazzoli · Infotech@Aerospace 2011 · 2011
In this paper, we present a new approach to stochastic scheduling for aircraft carriers, based on a queueing network model of the flight deck operations. The nodes of the network consist of the different stations which serve the aircraft, like refueling stations, catapults, and landing strip. Each server in this network follows a certain policy for servicing the aircraft, and then decides which of its adjacent nodes the aircraft will visit next. The focus of the problem is to find methods for generating good scheduling and routing policies. We find the optimal scheduling policy from a restricted subset of the set of all policies, namely linear index based policies. The parameters chosen to compute the indices are those that best differentiate the aircraft, and a linear index based structure is selected to make the policy easily implementable by the human operators. The optimal policy is then found by simulation based optimization. We finally obtain a policy that is implementable by the decision makers at each node making it, in a sense, a distributed policy. Moreover, we show how this framework can robustly deal with system failures, such as unplanned shutdowns or task execution failures that occur in the system. One interesting feature of this formulation is that it can be applied to a variety of problems like job-shop scheduling, and factory production.