Resource Optimization of a Software Development Teams Working with Kanban

Kanban has been used as a framework in project management for software development for almost two decades. Rooting to lean principles, Kanban helps teams to decrease cycle time, increase efficiency, eliminate waste through a number of techniques including value stream mapping, work-in-progress (WIP) limits, process analysis, analyzing and identifying wastes and lastly headcount planning. This study investigates optimization of the value delivered by a software development team working with Kanban via simulations and a clearing function integrated to a decision tree regression. The simulations in this study include multiple work item types with value, priorities and block scenarios. The resources conducting the work have separate skill sets and costs based on the skills. The block scenario indicates any work item to be blocked for some time based on a probability at the processing station. In order to find an optimum set of resources taking generated value and cost of resources into account, a decision tree regression model is compared to two benchmark models