Research On Software Cost Estimation Based On Vague Set

On account of the diversification and complexity of software projects,project manager can't work out the development budget effectively,unable to allocate resources reasonably,and finally lead to software project budget overruns,which affects the normal development of the project.Research shows that the instability of the initial stage of the software project and the inaccurate cost estimation are the main reasons for the failure of software project.In the cause of the failure of software projects,frequent changes of requirements are inevitable.Therefore,it is more necessary to carry out reasonable cost estimation at the initial stage of the project.Analogy-based Estimation is a common software cost estimation method based on historical cases.This method does not depend on the experts' judgement in cost estimation,and it can estimate the budget of the project more objectively.However,in the process of software project measurement,project managers often use fuzzy language to express the project.The traditional distance based similarity method cannot effectively quantify such data.At the same time,the relationship between different cost factors in software cost data is more complex,and the traditional analogy estimation method cannot effectively deal with such information.Therefore,it is necessary to introduce the Vague set to deal with the software cost data,and to improve the traditional analogy estimation method using the advantages of the fuzzy information processing.Vague set theory as an extension of the fuzzy set,can provide support for membership at the same time,against the membership degree and unknown information,said a more comprehensive information,which has been widely applied in the field of decision making,better to deal with fuzzy information,to distinguish between different data.Development time is shorter due to the Vague set theory,some standardization axiom is not perfect enough,similarity measurement method cannot completely meet the requirement of existing,thus need to the existing set of Vague similarity measurement method for research and its application in software cost estimation.Based on the analysis of existing software cost methods,this paper studies the following aspects:1.Summary of the last ten years,the software cost estimation research results in the field of domestic and foreign scholars,and based on previous literature research on the impact factors of software project cost,the staff factors,project environmental factors,itself factors and progress factors as the division basis,construct a reasonable estimate for the software cost index system.2.Analysis the flaws of the traditional method for estimating the analogy,in the process of measurement of software project,use Vague values to represent the data for a software project and make it can express the traditional method based on distance similarity can't effective quantitative representation of the degree of fuzzy information,to determine the optimal size between different project properties.3.Study on Vague set theory,in view of the existing shortcomings in the course of Vague set similarity measurement method was improved,the contains information support and opposition and unknown degrees under the premise of adjustment parameters,propose a new similarity measure of Vague sets method,and the theoretical proof and verification.4.Based on the traditional analogy estimation method as the basic framework,the proposed method of Vague set similarity measurement is used as the similarity function of software project to reduce the estimation error.In order to avoid one-sided and subjectively determining the weight of software project attributes,the entropy method of Vague entropy weight method is applied to determine the attribute weight which is more objective than other methods;Finally,it proves that this method can effectively reduce software cost estimation error and improve estimation accuracy.