There is no universal best formulation
A predominantly bio-based rigid package may need sufficient stiffness, low mass, controlled raw-material cost and a plausible processing window. Increasing one component can improve one objective while worsening another. The answer also changes with the selected grade, supplier price and manufacturing route.
A single headline score can conceal these tensions. It embeds preferences that may be appropriate for one project and wrong for another. Engineering teams need to see the alternatives behind the score.
Separate mandatory limits from preferences
Composition totals and permitted component ranges are hard constraints: a candidate either satisfies them or it does not. A desired cost or bio-based threshold may instead be a soft target, depending on the project. Keeping the two categories distinct avoids rejecting useful candidates for negotiable reasons.
The same discipline applies to missing evidence. If a required component property is absent, the candidate should not receive a fabricated substitute. It may remain visible as unevaluated, with the missing input stated clearly.
Read trade-offs as a portfolio
A multi-objective study can reveal a set of candidates for which improving one objective requires sacrificing another. That set is more informative than a single winner because it shows where the project can move when commercial or technical priorities change.
A good shortlist also preserves diversity. Eight nearly identical formulations teach less than eight candidates spanning meaningful regions of the design space. Diversity makes a validation campaign more informative and less vulnerable to a narrow computational optimum.
The physical experiment closes the loop
Early engineering estimates cannot establish actual tensile strength, impact behaviour, permeability, thermoformability, compostability or regulatory compliance. Those properties require suitable validated models or physical evidence under relevant conditions.
The computational result is therefore a research priority. After compounding and testing, measured values should be compared with the corresponding estimates without overwriting either record. The difference is itself valuable evidence.
KEY POINT
The right output is a transparent portfolio of compromises, followed by physical validation—not a formula presented as certain.