A systematic approach for prioritizing software requirements

Abstract: In most commercialdevelopment projects, there are more candidate requirements subjectto implementation than available time and resources allow for. Acarefully chosen set of requirements must therefore be selected forimplementation. A systematic approach for prioritizing candidaterequirements is a very useful means to provide necessary and usefulinput for the crucial selection decision.This thesis provides results from the development and applicationsof different approaches for prioritizing requirements in closecollaboration with Ericsson Radio Systems AB. A pairwise comparisonapproach for prioritizing requirements according to multiplecriteria has been developed and applied. To overcome the highnumber of comparisons that the approach often required in projectswith many requirements, different candidate approaches have beeninvestigated and applied for reducing the required effort. Anapproach for managing requirement interdependencies and theirimplications for the prioritizing approach has been developed. Asupport tool packaging the prioritizing approach and automatingmuch of the manual work in the approach has been developed andevaluated in practice.Qualitative results indicate that the proposed approach is aneffective means for selecting among candidate requirements, forallocating resources to them and for negotiating requirements. Theapproach further enables knowledge transfer and visualization,helps to establish consensus among project members and creates agood basis for decisions. Quantitative results indicate that therequirements actually selected for implementation have a profoundimpact on the final product. In several projects where requirementswere prioritized according to the criteria value for customer andcost of implementation, implementing the requirements whichoptimize the relation of value for customer to cost ofimplementation would reduce the development cost and developmenttime. Software systems with substantially the same value forcustomer can consequently be delivered with a reduction in cost andlead-time when the proposed prioritizing approach is deployedcarefully.