Research Of Effort Estimation For Software Project Development Based On COCOMO Ⅱ

In allusion to the deficiencies existing in current effort estimation of software project such as the low estimation accuracy, the complicated estimation method and so on, the thesis studies the construction of effort estimation of software project with efficient, high precision and easily operation. Base on the plentiful, reliable and precise experiment data accumulated by the software enterprises, the thesis mainly research on the follow five aspects.By extensive collection of the literatures and research on the present situation in effort estimation of software project at home and abroad, the relation between effort estimation and software project development is analyzed. And the deficiencies existing in the current effort estimation of mobile application software project as well as the reason for it is studied. Then the necessity to study of mobile application software project effort estimation and build accurate, universities, convenient effort estimation method for software project is discussed.As to the existing deficiencies of current software project effort estimation such as the low estimation accuracy, the complicated estimation method and so on, the achievement of the current effort estimation research is collected and problem as well as the reason of the existing effort estimation model is researched. And the operation and the performance of the existing estimation model is focal analyzed.To solve the problem about effort estimation of mobile application software project development, the method of selecting effort estimation model is analyzed. And the method of mobile application software effort estimation is presented based on COCOMO Ⅱ model. Under the condition of demonstrating the necessity and the feasibility of the estimation method, the process of the effort estimation of mobile application software is discussed.Then the advantage and the limitation of applying the standard COCOMO Ⅱ model to estimate the effort is studied. As to the present situation in effort estimation of software project at home and abroad, the necessity to localize the standard COCOMO Ⅱ model is demonstrated and the rule of localization is presented. As a result, the localization step for economies factor of scale, cost driver factor and the calibration parameter is presented. Further more, effort estimation process of mobile application software based on localized COCOMOⅡ model is presented.Thus, for the demonstration of the presented effort estimation of software project based on COCOMO II model, the evaluation of criterion for the estimation model is discussed deeply and the strategy on evaluation of effort based on the evaluating indicators is put forward in this section. This paper gathers sample data of presented effort estimation of software project and compares the effects of evaluation based on COCOMO II model, BP neural network, and grey correlation analysis and demonstrates that evaluation based on COCOMO II model can reduce average error and raise the reliability.