WED 11:30 AM Parallel Session: Military Applications
Full Report: How to Deliver Multi-phase Software Development Projects: System Dynamics Simulation of Alternate Project Strategies, by Alan McLucas
The main objective of the first paper, presented by Alan McLucas, was to examine project management strategies in the development of military software. The presentation focused on the allocation of resources throughout the development process, particularly in response to defect detection and for making any adjustments. It was noted that linear heuristics for the assessment of project management requirements was inappropriate for change work in software development due to rework feedback loops in action. The presentation focused on the four stages of code development and the dynamic allocations of resources highlighting that adding additional resources at the end of the project was not an effective strategy. It was demonstrated that a preferable strategy was to ‘front load’ the allocation resources at the beginning of the project, arguing that effectiveness in identifying defects in the early phases of a project improved the likelihood of successfully delivering subsequent phases. Question: Given that the importance is optimising the time of the whole project as opposed to phases, what happens when one stage becomes too much of a focus? Answer: The strategy demonstrated to improve completion of the project is to allocate resources early in the project; this requires that there are sufficient people trained prior.
Investigating the Effectiveness of Various Crowd Confrontation Strategies Using Vensim and the Phoenix Integration Suite, by Ivan Taylor, Luminita Stemate, Guillaume Gagne, and Anissa Frini
The second presentation discussed ways of evaluating the effectiveness of non-lethal weapons in crowd control situations, the results of a three-year defence research project in Canada. Different types of crowd members were presented, categorised by their crowd behaviour and their effect on the aggressiveness level of the crowd as a whole. Instigators/leaders in the crowd tried to make the crowd more aggressive (to a desired level) whilst the control forces tried to bring down the level of aggression. Four sub-models have been created with people classified by their social categories/demographics. Aggressiveness levels of the crowd were determined from actual data using the Vensim calibration tool. The effect of control force interventions, including a combination of tactics, was assessed as an integer optimisation problem within the Phoenix Integration suite, the method being based upon a Darwin genetic algorithm. Question 1: Does the model consider geographical allocations? Answer 1: No, though an agent-based model is being considered to represent this factor. Question 2: How scalable is the model? Answer 2: Easy with system dynamics, more difficult as an agent-based model because of computing power. Question 3: What about design in city structure as a factor in crowd control, such as wide avenues? Answer 3: Inclusion of this effect is difficult; it is complex problem e.g. the inclusion of escape routes for forces.
Evaluating Strategies for Controlling Attrition Rate of Canadian Air Force members, by Arif Mehmood and Jules Vanie
A human resource model of the Canadian Air Force staff was presented by Arif Mehmood and Jules Vanie in the third presentation. Many factors, such as work overload, imbalance between work and family life, and delays in training and promotions, were considered. Discussion of the utility of the model centred on its adoption by senior management, to aid which a Sable interface was developed and presented allowing easier adaption of the model parameters. Question 1: Have you done any policy analysis so far? Answer 1: The issue justifies this methodology as a basis for policy analysis; further analysis is needed. Personnel managers have stated that the human resource departments are sceptical about the system dynamics approach. Currently air force personnel are comparing the effects of retention strategies from other projects. Question 2: It appears that the attrition rate has trebled over the past 3 years. If the crunch has come, are they doing something? Answer 2: They are collecting data to analyse the situation; the issue is ‘what is to be done?’ In terms of implementing solutions there is no clear policy. The current state of the system can partly be attributed to an earlier mandated policy to decrease the numbers of the air force. Now that people are retiring, the effects of this policy are being realised. In addition, younger generations are more likely to move careers than older generations, resulting in high attrition rates; this is compounded with the lengthy training period required for a career in the air force. Question 3: Is the issue of convincing senior management regarding system dynamics methodology the quantification of soft variables? Answer 3: More so the difficulty in policy validation given the time frames involved in observing any effect. Question 4: Why did you develop a fancy interface if you worry that senior management is not convinced in the methodology? Answer 4: The reason we created the interface was to help the managers see how flexible and useful such a tool would be, and that, as such, there would be no need to train staff specifically. Also, interfaces meet the needs of the decision makers as this is what they are familiar with. Generals like interfaces.
James Rhys Kearney