Force generation plan automation for the royal australian navy

Abstract

The Royal Australian Navy's (RAN) Force Generation Plan (FGP) details the day-to-day operations of about 120 ships, crews, submarines, diving teams and air assets plus miscellaneous entities. The FGP is an annual directive concerned with scheduling the activities that are undertaken by the RAN units. Required activities are specified in an associated document called the Navy Activity Schedule (NAS). The NAS provides overarching guidance to RAN maritime operations planning staff on Navy's priorities and resource levels for operational commitments, preparedness levels and international and domestic engagement activities. The FGP is manually generated as a result of interactions and conferences involving Navy operations' authorities. A scheduling tool, the Fleet Activity Management Tool (FAMT), allows schedulers to coordinate unit assignments to activities while keeping track of other factors such as requirements for units to undergo maintenance and personnel to have rest and recreation. The scheduler makes heuristic choices based on 'business' rules dictated by policy documents. However, while FAMT can help to generate the FGP, the actual scheduling of units to NAS activities requires an operator to make the assignment. Manual scheduling can be time consuming and is not guaranteed to cover all the required activities in the most efficient manner. Typically, much iteration of the schedules occurs. This paper describes a simple decision support system for scheduling RAN fleet units using automation. The paper starts by describing the employment allocation problem and the constraints conditioning the generation of an appropriate overall schedule. Then it focuses on describing the heuristic approach of the solver using a simulated annealing algorithm, together with its flexible interaction with the planning tool FAMT. The optimisation model developed here offers schedulers the advantage of considering many candidate schedules to determine the best schedule. Apart from the obvious opportunities for labour reduction, automation also offers the advantage of providing a consistent repeatable process with robust evaluation criteria as to the quality of a FGP. In comparison to an actual FGP, the FGP generator demonstrated an improvement in the allocation of units to the NAS by about 30% on key fleet management measures, with commensurate improvements to training opportunities and preparedness. Trial runs of the FGP generator consistently found satisfactory unit allocation schedules for difficult pathological scenarios. This research shows that automation of the FGP can lead to substantial improvements in efficiency, while maintaining satisfactory unit programmes, and ensuring that RAN fleet units can meet their on-going and emerging commitments. The solver has been integrated into FAMT and is available to schedulers. The paper notes some avenues for improvement of both FAMT and the FGP solver and the method's applicability to wider defence resource allocation problems.