In development, it is difficult to foresee all factors and model structures during the initial experimental stage. After reviewing experimental results, it may be necessary to revise the search range, add interaction or squared terms, or introduce new explanatory variables. On the other hand, discarding acquired experimental data and restarting the experiment from scratch is a significant burden in terms of time and cost. JMP's extension plan is a function that allows you to design additional experimental points necessary for developing the model, starting from existing data. This presentation will introduce the concept and application procedure of extension plans, which allow you to develop the model step by step while utilizing existing data based on insights gained from initial experiments. Specifically, it will show the flow of generating additional experimental points and updating the model using JMP, assuming situations that are likely to be encountered in practice, such as adding interaction or squared terms, revising the search range, and introducing new explanatory variables for results that cannot be fully explained by the initial experiment alone. We propose a practical method for connecting experimental results to the next plan and deepening technical knowledge within limited experimental resources.
