| With the continuous improvement of the informatization level of the aerospace system,the proportion of software in the aerospace system is increasing.The quality of aerospace software is related to the success of aerospace missions and the effectiveness of the equipment in orbit.In order to describe whether the quality of aerospace software products meet the requirements of users objectively,it is necessary to make a reasonable and scientific evaluation of the aerospace software quality.Therefore,it’s of great significance to carry out corresponding research of aerospace software quality models and evaluation methods.The existing quality models tend to focus on general adaptability to different fields and different types of software,so these models may not be complete and comprehensive when used to evaluate aerospace software with strong domain characteristics,in addition,the existing quality models do not fully consider the differences in the measurement of the same quality characteristics at different stages of the software development process.The fuzzy comprehensive evaluation method can make a conscious comprehensive evaluation.However,there are some problems in this method such as the loss of valid information and the possible failure of the maximum membership criterion.In traditional software quality evaluation methods,the weight of each indicator element is generally set based on expert experience,and qualitative information is mostly quantified through human subjective judgment,so there is the risk of subjective arbitrariness and idea uncertainty,in addition,the traditional evaluation methods usually ignore the nonlinear relationship between each measurement index and software quality,and lack self-learning ability,so the evaluation results obtained are often difficult to truly reflect the real quality of software.In view of the above problems,on the basis of the summary of traditional software quality models and evaluation methods,research work on the quality evaluation methods of aerospace software is carried out in this paper based on the characteristics of aerospace software.The main contributions and research results of this paper are as follows:1.Based on the analysis of the existing quality models,according to the characteristics of aerospace software,an aerospace software quality evaluation model of full development life cycle is constructed.The quality characteristics and quality sub-characteristics in the model are described in detail,and the metrics for each quality characteristic in different stages of the software development process are designed.The aerospace software quality model based on hierarchy can fulfil aerospace software quality evaluation requirements.2.An improved fuzzy comprehensive evaluation method for aerospace software quality based on confidence is proposed.On the basis of designing effectiveness judgment indicator,before the maximum membership criterion is used to get the evaluation result,the validity of the criterion is judged first.Once the criterion is inefficient or ineffective,the confidence criterion is used to obtain the final evaluation result.Through the application of actual calculation examples,it is proved that the proposed method is effective and feasible.3.The dynamic fuzzy neural network is introduced to implement the intelligent evaluation of the aerospace software quality.Besides,an aerospace software quality evaluation method based on SCS-D-FNN is proposed.4.The optimization method of the dynamic fuzzy neural network is studied,and the optimization algorithm is used to optimize the network,besides,the optimized dynamic fuzzy neural network is used to evaluate the aerospace software quality.An improved version(SCS)with Stud crossover operator is proposed,and the SCS optimization algorithm is used to optimize the dynamic fuzzy neural network to improve the evaluation performance of the network.The experimental results prove that the proposed SCS algorithm has better optimization performance.The experimental verification of the evaluation performance of the SCS-D-FNN model proves that the proposed SCS-D-FNN model has good evaluation performance,and has objective and practical advantages. |
