Selecting the appropriate design type for a Design of Experiments (DOE) campaign requires careful consideration of various constraints and factors. Firstly, consider the aims of your DOE campaign. Are you looking to explore, optimise, or screen? Do you need to make highly accurate predictions, or just quick decisions? DOE can help with all of these things, but first, you need to select the right design type and outline the DOE campaign structure.

Make sure to establish what constraints you might have, and take into account the prior knowledge and expected outcomes of the DOE.

Research Objectives: Clearly define your research objectives. If the goal is to identify the most important factors quickly, a screening design might be appropriate. If precision in estimating specific factor effects is vital, an optimal design might be more suitable.

Predictive vs screening vs explorative models: If you plan to develop predictive models based on the experimental data, the choice of design might impact the model's accuracy and generalizability.

Complexity of Interactions: Consider the complexity of interactions between factors. Full-factorial designs excel at exploring interactions comprehensively, while optimal designs might be better for detecting specific interactions of interest.

Trade-Offs: Recognise the trade-offs between exploration and precision. For example, space-filling designs are comprehensive but require more resources, while screening designs are efficient, but might sacrifice precision.

Prior Knowledge: Existing knowledge about the system can directly and indirectly guide your design selection. If there is prior information that suggests certain factors are critical, then use it to influence the choice of factors. For example, if you expect that certain interactions between factors will definitely play a role in the behaviour of your system, then make sure you select a design that will allow you to investigate interactions effectively.

The factor type (e.g. continuous or categorical), and their levels: Factor types and levels may constrain the type of designs you can use.

Number of Factors: The number of factors being investigated is crucial. Screening designs are suitable when dealing with a large number of factors, while full-factorial designs become increasingly large and complex as the number of factors grows.

Experimental Constraints: Consider any physical constraints imposed by the experimental set-up or equipment. Some designs might not be feasible due to equipment limitations of factor types (i.e. continuous vs categorical factors).

Resource Availability: The availability of time, budget, materials, and personnel, significantly influences the choice of design. More resource-intensive designs, such as full-factorial or space-filling, may require larger budgets and take longer to execute.

Experiment Replicates: The number of replicates (repeated experiments) for each factor level should be considered. Think about the expected noise in your experimental data, the anticipated effect size of your factors, and balance this with constraints to your run number.

In essence, the decision regarding the appropriate DOE design is driven by an evaluation of the research goals, available resources, and the complexity of the experimental system. By considering these constraints and factors, scientists can make an informed choice that aligns with their objectives and maximizes the value of their experimental efforts.

To learn about the different DOE design types supported in Synthace and which one might be right for you, click here.