Pareto optimization

As explained in Optimization Algorithms: AI techniques for design, planning, and control problems

A multi-objective optimization technique that seeks multiple nondominated trade-off solutions rather than one scalar objective optimum.

Pareto optimization handles conflicting objectives by finding multiple trade-off solutions instead of forcing every goal into one weighted score. It reduces the alternatives to a Pareto frontier: an optimal set of nondominated solutions from which a final choice can be made using higher-level information. section 1.3.2 section 8.9

Use the approach when:

• Objectives conflict and assigning subjective weights would obscure the decision. section 1.3.2
• Several nondominated alternatives should remain available for strategic decision-making. section 1.3.2
• A genetic algorithm is being adapted for multi-objective optimization; crossover and mutation can help explore candidate solutions. section 8.9

Electric vehicle design

Choose wheel size, motor power, and battery capacity while maximizing EPA driving range and minimizing acceleration time. A larger battery can extend range but add weight, increasing acceleration time. Rather than selecting weights for the two objectives in advance, Pareto optimization identifies the available range–acceleration trade-offs. section 1.3.2

Figure 1.6 Electric vehicle design problem for maximizing EPA range and minimizing acceleration time

Inspect the electric-vehicle design figure to compare the two competing objectives. section 1.3.2

How Optimization Algorithms: AI techniques for design, planning, and control problems covers it

How deep it goes

Defines nondominated solutions and contrasts them with weighted scalarization, but defers detailed algorithms to genetic algorithm variants.

Examples

Uses electric-vehicle design with battery capacity, EPA range, and acceleration time as a conflicting-objective example.

What it assumes

Assumes objective functions and conflicting maximization or minimization goals, while contrasting them with weighted scalarization.

What you can do after

Evaluate trade-offs and retain nondominated alternatives before making a higher-level strategic choice.

How it connects

Connects multi-objective optimization with objective functions, preference-based optimization, genetic algorithms, and decision-making.

liveBookmanning.com

Read Optimization Algorithms: AI techniques for design, planning, and control problems on liveBook Get the book at manning.com

Related concepts

• Pareto frontierproduces · A Pareto frontier is the set of feasible solutions that are not dominated by any other solution across the objectives being optimized.
• multi-dimensional optimization problemkind of · A multi-objective optimization problem involves jointly optimizing two or more objective functions, which may conflict and require balancing trade-offs.

This explainer was generated by Manning's knowledge graph from the text of Optimization Algorithms: AI techniques for design, planning, and control problems. The section links point to the parts of the book it draws on.