University of Nottingham Malaysia
School of Computer Science
     
  

Dynamic Risk Management Framework for Software Projects

Risk management techniques and tools are evolving through continuous research. Many researchers are contributing toward the topic dealing with interesting question and scenarios of risk management framework including the management, assessment, identification, prevention and mitigation of risk factors. In the development of software projects cost estimation is the single largest factor of risk and uncertainty. Software cost estimation is sometimes referred as effort estimation since the effort translates into cost. Much of the software risk management research work has focused around uncertainties involved in the cost and effort estimations.

Research work on software risk management and assessment has focused on a static view of the risk factors and assume that changes through passage of time will not have an impact on risk factors. The current risk management framework does not deploy feedback loops and mid course interventions based on the changing environment to counter the possible impact of the risk factors. It is much needed to use stochastic processes to bring the time varying element into the risk management and assessment framework hence making it more dynamic to the changing environment.

Static risk analysis assumes that the probability distribution and its related parameters i.e. mean, median, dispersion, skewness are fixed and remain fixed for the duration of the project. It does not capture the time varying element or in other words the uncertainty around the distribution parameters and hence presents a two dimensional view of the probability distribution and of the problem. These models capture the uncertainty of the risk factors through the shape of the suggested probability distribution.

A stochastic process on the other hand captures the time changing uncertainty and incorporates it into the analysis process and assumes that the shape of a suggested probability distribution could change overtime due the uncertainty involve in the distribution parameters. Stochastic process or random process is the counter part of the deterministic process and instead of dealing with one possibility it encompass some uncertainty about the possibilities or probability distributions. Over a period of time a random process can take any path thus introduces a sense of time in the probability distribution. Stochastic process amount to sequences of random variables called time series and provides a three dimensional view of the probability distribution.

In a static forecasting environment there is no way to quantify the variability of the possible outcomes or to easily see the full depth and breadth of the possible outcomes. Thus dynamic framework model makes it possible to describe critical assumptions and their combined implications in terms of possible outcomes. As a most evolutionary step dynamic framework modelling incorporates feedback loops and management intervention decisions into the model.

Dynamic risk management framework is an evolving standard managed by the Casualty Actuarial Society. In the simplest terms dynamic risk model can be used to analyze variables under consideration through stochastic simulation of the future values of those variables. In the field of economics the dynamic risk management models are widely used from stock markets, Insurance companies, banks and money lenders.

The dynamic risk management framework differs with the static risk management in the way that dynamic risk assessment involves the simulation of the probabilities of risk events under uncertainty and it deploys feedback loops as preventative action should a risk event appears. The preventative feedback actions are followed by new set of probability assessments and preventative actions. The preventative actions should change the perceived probabilities of future events so the risk assessment taken at different times may estimate different quantities. The natural risk events stays the same what changes is the probability assessment due to the natural variability and preventative actions taken at different times.

 

 

School of Computer Science

University of Nottingham Malaysia
Jalan Broga, 43500 Semenyih
Selangor Darul Ehsan
Malaysia

telephone: +6 (03) 8924 8000
fax: +6 (03) 8924 8018

Make an enquiry