Process Strategies for Large Scale Software Development – Simulation using Systems Dynamics

Peter Henderson and Yvonne Howard
Department of Electronics and Computer Science
University of Southampton
SO17 1BJ, UK

Today's commercial environment demands fast responses to new needs. Producers of large scale software recognise that software evolves and that advanced process techniques must be used to maintain competitive responsiveness. CMPM, the Cellular Manufacturing Process Model is an advanced component based process strategy which uses concurrency and distribution to reduce cycle times. In CMPM networks of semi-autonomous cells co-operate to produce a complex large scale system. The model views development as a manufacturing activity where systems are built from components, which are a mixture of self built components, re-used components and from bought in components. The model is hierarchical, any component may be a product of others.

Software producers need predictability when competitive advantage demands a short time to market. Predicting the cost, quality and schedule outcome of CMPM depends upon the behaviour within cell (intra cell) and the co-operative behaviour between cells (inter cell) in a dynamic environment.

Evaluating the effects of CMPM on cycle times and predictability is an active research area with the support of our industrial partners, ICL. The aim of the research is to provide a simulation based tool for designing and dynamically controlling CMPM processes.

This paper examines some of the issues that affect the ability of cells to achieve their targets. We explain how we are using Systems Dynamics modelling and simulation to develop our understanding of both inter cell and intra cell behaviour and to provide evidence of the benefits of CMPMand identify the control points that predict performance.

Keywords : systems dynamics, process model, cycle-time acceleration

<BACK TO PROGRAM SCHEDULE> <HOME>