A quantum-based approach to error in software development effort estimation

Abstract

Error in Software Development Effort Estimation is often investigated in order to adjust effort estimation technique results. The aim of this paper is to propose a quantum-based approach in order to deal with model error whatever the effort estimation technique used. To achieve this objective, we explored the Quantum Theory and proposed a model based on an analogy with the infinite potential well problem. In order to evaluate the performance of our approach, we assess the Classical Analogy based effort estimation technique on the COCOMO'81 dataset. The main findings are: (1) the proposed approach confidence intervals are close to the theoretical ones and (2) the Sine-squared distribution of effort represents an interesting alternative to the widely used Gaussian distribution since both distributions are statically comparable. These results give an interesting track for further researches and investigations.