Design of Experiments Iteration Agent

Optimizing a process means testing combinations of temperature, time, concentration and other factors, and each lab run costs time and material. This agent proposes an experimental design that fits your factor ranges and run budget. The scientist approves each round before it goes to the lab. When results come back, the agent fits a model, checks its quality and residuals, and identifies which factors matter. Runs with a known lab error are excluded with a note. If the model is weak or the optimum sits at the edge of the tested range, it proposes the next round, such as extending a range or adding center points to test curvature. It stops when the optimum is found with enough confidence or the run budget is used. It then writes the optimum conditions with confidence ranges. Edge case: a run with a known pipetting error is excluded and noted, not silently dropped.