Objective Endpoints

Objectives in globalMOO define the goals for optimization trials. They specify target values and constraints that the optimization process aims to achieve.

Available Endpoints

• Load Objectives - Define objectives for a trial

Objective Types

globalMOO supports various types of objectives:

1. Target-Based Objectives

   • exact - Match target value exactly (within L1 norm)

   • percent - Within a percentage of target value

   • value - Within an absolute value of target

2. Constraint-Based Objectives

   • lessthan - Less than target value

   • lessthan_equal - Less than or equal to target value

   • greaterthan - Greater than target value

   • greaterthan_equal - Greater than or equal to target value

3. Optimization Objectives

   • minimize - Find the minimum possible value

   • maximize - Find the maximum possible value

Note: All objective type strings should be in lowercase in API requests.

Bounds Handling

For percent and value objective types, bounds must be specified:

• minimumBounds - Lower bound of acceptable deviation (required)

• maximumBounds - Upper bound of acceptable deviation (required)

For exact type objectives, bounds should be set to 0.0, and the desiredL1Norm parameter controls convergence.

For constraint-based objectives (lessthan, lessthan_equal, greaterthan, greaterthan_equal), bounds are ignored.

For optimization objectives (minimize, maximize), bounds can be used to limit the search space.

Common Use Cases

1. Multi-objective optimization with mixed constraints

2. Target matching with acceptable tolerances

3. Minimization/maximization with boundary conditions

4. Constraint satisfaction problems