Neural combinatorial optimization

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

A research area focused on applying machine-learning models and algorithms to combinatorial optimization problems.

Neural combinatorial optimization applies machine-learning models and algorithms to discrete, high-dimensional optimization problems, often converting graph data into representations that support solution generation. section 11.10 section 11.5

Use machine learning in three complementary ways:

• Generate solutions end to end from an input problem instance. section 11.5
• Configure an optimization algorithm by learning parameters such as step size, cooling schedule, mutation variation, or swarm coefficients. section 11.5
• Guide an optimization algorithm repeatedly by using the model’s decisions about its current state. section 11.10 section 11.5

Pointer networks for routing

A pointer network, a sequence-to-sequence model with attention for variable-sized inputs, can be trained on m points and validated on n points for a Euclidean planar symmetric traveling-salesperson problem. section 11.10 section 11.5

Figure 11.14 Machine learning (ML) for combinatorial optimization (CO) problems

A visual map of the three ways machine learning can generate, configure, or guide combinatorial optimization. section 11.5

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

How deep it goes

Distinguishes three application patterns and gives parameter examples, but stops short of implementation details or formal analysis.

Examples

Uses pointer networks for Euclidean planar symmetric traveling-salesperson problems and parameter learning for simulated annealing, genetic algorithms, PSO, and ACO.

What you can do after

Apply it by selecting direct solution generation, learned algorithm configuration, or repeated model-guided optimization.

How it connects

Connects sequence-to-sequence models, attention mechanisms, pointer networks, supervised learning, unsupervised learning, reinforcement learning, and optimization algorithms.

liveBookmanning.com

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

Related concepts

• machine learninguses · Machine learning is a field of computational methods in which systems learn patterns or relationships from data to make predictions, decisions, or other outputs.

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.