Multi-Objective Optimization: Choosing a Strategy
When your experiment has more than one objective, the platform asks you to choose a multi-objective strategy. This guide helps you pick the right one.
Discover vs Prioritize
The first choice is between two modes:
Discover | "What are my options?" — Explore trade-offs and see a range of optimal solutions. |
Prioritize | "What is the best single solution?" — You already know which objectives matter most. |
Three Strategies
| Pareto Front | Weighted Sum | Hierarchy |
Mode | Discover | Prioritize | Prioritize |
Output | Multiple trade-off solutions | One best solution | One best solution |
You set | Nothing (automatic) | Weight % per objective | Priority order + tolerance % |
Best for | Early exploration | Known importance, comparable objectives | Strict ranking, non-comparable objectives |
Read the detail article for each strategy:
Pareto Front Optimization — explore optimal trade-offs between objectives
Weighted Sum — assign percentage weights to each objective
Hierarchy — rank objectives in strict priority order with tolerances
Which Strategy Should I Use?
Question 1: Are all your objectives equally important?
Yes, equally important — Use Pareto Front. Explore the full trade-off front between objectives.
No, some matter more — Use Prioritize (Hierarchy). This is the recommended default — rank your objectives by importance and set tolerances. The optimizer focuses on the top priority first.
Question 2: If using Prioritize, can you express importance as exact percentages (e.g. 70/30)?
Yes — You may also consider Weighted Sum as an alternative. Assign weights and get one best solution.
No — Stay with Prioritize (Hierarchy). It handles non-comparable objectives naturally.
Quick Tips
Prioritize (Hierarchy) is the recommended default whenever any objective is more important than the others.
Pareto works best with 2-3 objectives and only when all are equally important. For more objectives, use Prioritize or Weighted Sum.
An experiment can have up to 10 objectives, each set to Maximize, Minimize, or Target.
Adding a 2nd objective defaults to Pareto Front. You can switch anytime.
All objectives in an experiment share the same strategy.