A DMAIC- based approach for systematic testing of an automated prescription-filling system.

摘要

Information about the quality of a system is of utmost importance to all the stakeholders. Testing is the process by which the quality of the system can be determined. The objective of testing is to find errors. The final system test is the last line of defense before the machine is shipped to the customer. Hence, the incidence of detection of defects has to be maximized during final system testing. Achieving wider test coverage necessitates a meticulously developed test plan. This research aims at developing such a test plan for an automated prescription-filling system.;A DMAIC (Define, Measure, Analyze, Improve, Control) based methodology is employed in this research. The key component of the final system test is the duration to which the machine needs to be tested so that maximum coverage can be obtained with optimal resources. The duration or the number of cycles to which the machine needs to be tested primarily depends on the desired quality. A statistical approach to the determination of target quality of the system by the estimation of target quality levels of sub-systems is presented. Three methods by which the target quality levels of sub-systems can be estimated are described. These include data from new design testing, past in-house tests and data obtained from the field machines. The importance of beta testing during the new design implementation is also addressed.;The methodology is implemented for developing the test plan for the final system testing of the automated prescription-filling system. Statistical calculations employed in the estimation of the cycles for the final system test are presented in detail. The results indicated that the sample size varies widely with confidence levels and precision levels. These values should be chosen based on practical considerations such as the time available for final system testing. In conclusion, this research emphasizes the importance of a test plan, i.e., formal testing strategy for the final system testing. The need for formal statistical estimation of the duration of testing for achieving maximum test coverage with optimal resources is substantiated.