Experimental Design s for Development Tests in Distributed Joint Environment s

TLDR: This paper explains some basics of fractional factorial split plot designs and shows how to use the concept of partial design resolution to produce better designs based on specific test objectives.

Abstract: The c haracteristics and objectives of development tests in distribut ed live, virtual, and constructive joint environments favor fractional factorial split -plot experimental designs . Development tes ts conducted in joint environments are key components of the Department of Defense transformational planning initiative to test as we fight . Joint operations have become the mainstay of war fighting. But t he only practical way to assemble joint environment s for testing is to link various live entities , virtual operator - or hardware -in -the loop simulations , and constructive simulat ion s ( simulated people operating simulated systems )-- all located at many geographically -dispersed test facilities . Such distributed tests are expensive and difficult to execute, resulting in many practical limitations that restrict the ability to completel y randomize trials. Objective s of development tests in joint environments include evaluating design alternative s based on joint mission effectiveness and assessing the robustness of design alternatives to uncontrollable sources of variation . These objecti ves and restrictions lead to fractional factorial designs with split plot components. This paper explains some basics of fractional factorial split plot designs and shows how to use the concept of partial design resolution to produce better designs based o n specific test objectives. An example using air -launched data link weapon s is used to illustrate design considerations and analysis concepts.