Performance evaluation of 3DCAD systems based on unified automatic test, cloud model and variable weight AHP

Abstract

Performance, typically measured in terms of accuracy, efficiency and resource occupation of various modeling operations, is a key factor for assessing the quality of 3DCAD systems, hence it is necessary to scientifically and impartially evaluate the performance of different 3DCAD systems for CAD companies to understand the gap between their systems and leading ones. In view of these considerations, a performance evaluation framework is proposed for 3DCAD systems based on unified automatic test, cloud model and variable weight analytic hierarchy process (AHP). First, an automatic test tool and a series of test cases are developed for the unified automatic performance test of various 3DCAD systems so that the test data can be efficiently achieved. In the acquisition of the performance test data, the fluctuation of hardware performance and the random resource allocation of operating systems will cause the uncertainty of 3DCAD systems itself and the instability of the test environment. To accurately modeling these uncertainties, a novel self-subtracted backward cloud generator (SSBCG) is proposed, based on which the performance indexes of 3DCAD systems can be characterized as cloud models and the uncertainty of system performance is distinguished from the external instability of the test environment by the entropy and hyper entropy of cloud models. Additionally, a variable weight AHP considering both the credibility and relative importance of the performance indexes is put forward to calculate the overall performance score of 3DCAD systems. Finally, the feasibility, effectiveness and engineering practicability of the proposed performance evaluation framework for 3DCAD systems is verified through a case study.