Computer Simulation and Markov Chain Modelling for HRM in the Belgian Defence

We present a set of new military manpower planning tools that were developed by the Royal Military Academy (RMA) in Brussels in close cooperation with the Directorate General Human Resources (DGHR) of the Belgian Defence between 2006 and 2008. The aim of this project was the development of a coherent manpower model and an associated manpower planning toolbox that can be used by the military manpower planners of the DGHR for short, medium and long term manpower planning, and this for the entire manpower population of the Belgian Defence, including the civilian personnel. We developed two separate models: a steady-state model and a dynamic model. The steady-state model is based on stochastic Markov chain theory and is intended for long term steady-state forecasting. It is in essence a theoretic model, as it doesn't take the actual current manpower situation into account. It shows the steady-state manpower situation that would be attained if a certain policy were to be maintained for a prolonged period of time. This manpower situation represents a steady-state situation: its age and rank structure can be maintained indefinitely, simply by continuing to apply the same policy. This is obviously a very desirable characteristic of a manpower distribution from a managerial point of view. If the manpower planner can identify the age and rank structure that fits the organisational framework required to fulfil the operational missions of the Belgian Defence, then we suggest that the steady-state manpower situation that fits those requirements corresponds with an “ideal” manpower population for the organisation. The second model is a dynamic model based on stochastic discrete event simulation. The main conceptual difference with the steady-state model is that this model is not theoretical but realistic: it interfaces with the day-to-day manpower management database and uses the actual current manpower distribution as its starting point. This model can be used for short, medium and long term manpower planning in a transient system state. Furthermore, the simulation methodology used gives the user the opportunity to assess the impact of the inherent stochasticity of the real-world environment. The dynamic model includes a search module, which uses a heuristic search algorithm in order to assist the manpower planner in identifying the most efficient manpower policy in a number of different scenarios. In the remainder of this paper we will further explain the architecture of both the steady-state and the dynamic model.