Dynamic Simulation of Software Workers and Task Completion

As more and more companies are trying to leverage crowd sourcing to reduce cost and time-to-market of software production, it is very important to better understand the dynamics of crowd sourced software development in order to fully realize its benefits and avoid hidden pitfalls. This paper presents a systems dynamic model for crowd sourced software development platforms based on data gathered from Top coder. The model is composed of three components: the worker sub-model, the task sub-model, and the interactions between the two sub-models. Results of simulations are compared with empirical data gathered from Top coder for validation. The major simulation results indicate that task uploading and software workers arrival in the platform follow typical exponential patterns, incentivizing the crowd developers to become more active will lead to more tasks completion and success, and tasks associated with higher awards generally receive few submissions.