Individual sliced minimum aberration designs for multi-platform experiments
摘要
Multi-platform experiments are widely applied in various fields, such as software testing, e-commerce promotional campaigns, and so on. Choosing optimal sliced designs for multi-platform experiments is an important task. A sliced design can be segmented into multiple sub-designs by the sliced factor, with each sub-design being allocated to a platform. In certain multi-platform experiments, the sliced factor is considerably more important than design factors and it is crucial to reliably estimate the effects of both design factors and the sliced factor. To address this requirement, we propose a new type of design, called individual sliced (IS) designs. Based on the sliced effect hierarchy principle (SEHP), we introduce the individual sliced minimum aberration (ISMA) criterion. By adhering to the ISMA criterion, ISMA designs can be selected as optimal IS designs to accurately estimate the effects of both sliced factors and design factors, particularly when sliced factor is more important than design factors, thereby providing methodological guidance for determining the factor level settings on each platform. This paper focuses on constructing ISMA designs and provides tables for ISMA designs with 32 and 64 runs for practical use. Additionally, we present a practical example to illustrate the application of ISMA designs in guiding promotional strategies on e-commerce platforms.