A Capability-Based Approach to Software Process ModellingValentina Plekhanova Despite recognition that project resource capabilities are an important factor in software production, little research exists to support people experiences and project resource capabilities/compatibilities. This paper presents a capability-based approach to stochastic software process modelling. Software process modelling is represented by a project schedule simulation which is considered as a combination of discrete event simulation and system dynamics modelling. In a model, the feedback path can be the reworking a task or any sequence of project activities (i.e. step, phase, period or stage). Since the number of loops is a random variable relevant to project resource capabilities, an integration of process simulation and system dynamics approach is used for stochastic software process modelling. A capability-based approach provides detailed process analyses, resource utilisation, and how resources (people/team) fit a project not only at each discrete point in time, but also for each feedback path. Monitoring of performance, feedback, and regular consideration of performance provides powerful stimuli for change in order to improve product quality. A measure of performance in scheduling problem defines a set of identical schedules, a set of equivalence classes of schedules, priorities/preferences among these schedules and provides the determination of an optimal schedule. This ensures a simulation of a sequence of a random number of optimal schedules, which represents the evolution of the schedule, with feedback effect over time. This approach addresses continuous optimisation of software development process in order to develop high quality software on time and within budget; at a desired performance/technology level while utilising the available resources effectively and efficiently.
|