Tutorial

Genetic Programming and Machine Learning for Scheduling

The IEEE World Congress on Computational Intelligence (WCCI 2026)/IEEE Congress on Evolutionary Computation (CEC 2026), 21-26 June 2026, Maastricht, the Netherlands

Tutorial Abstract

Scheduling is an important optimisation problem that reflects the practical and challenging issues in real-world scheduling applications such as order picking in warehouses, the manufacturing industry and grid/cloud computing. Job shop scheduling (JSS) is a typical scheduling problem, which covers a full range of topics and tasks including static JSS, dynamic JSS, flexible JSS, dynamic flexible JSS, from basic research to a huge number of realworld industrial applications.

Instead of manually designing scheduling heuristics and algorithms for each problem, we can use machine learning and hyper-heuristics to automatically learn effective scheduling heuristics from low-level heuristics, characteristics of scheduling problems, and dynamic information from production environments. Among the techniques studied and applied within the research field of JSS, genetic programming (GP), a powerful evolutionary machine learning technique, has been successfully used to learn scheduling heuristics for JSS, especially for dynamic JSS. Although GP has shown its advantage in learning scheduling heuristics for JSS, GP still has several limitations for handling JSS such as high computational cost and large search space.

This tutorial will provide a comprehensive introduction to evolutionary machine learning techniques for JSS. This tutorial will cover different types of (advanced) evolutionary machine learning approaches for JSS. From this tutorial, you are expected to get familiar with evolutionary machine learning in four aspects. First, you will learn the definition of hyper-heuristic learning with a comparison of heuristic learning. Second, the details of JSS (e.g., static, dynamic, flexible JSS) will be given. Third, how to use GP as hyper-heuristic approaches to learn heuristics for JSS will be introduced with examples. Last, this tutorial will show how to use advanced machine learning techniques such as feature selection, surrogate and multitask learning with GP to JSS. All the techniques mentioned will be introduced with promising results.

Outline

This tutorial contains seven parts:

1. Different types of scheduling and their applications

   1) Scheduling and its applications
   2) General introduction of job shop scheduling
   3) Static vs dynamic job shop scheduling
   4) Flexible vs non-flexible job shop scheduling
   5) The similarities and differences between different types of job shop scheduling

2. Evolutionary machine learning and genetic programming

   1) Hyper-heuristic learning
   2) Representation of genetic programming
   3) Evaluation of genetic programming
   4) Parent selection in genetic programming
   5) Genetic operators (i.e., crossover, mutation, and reproduction) of genetic programming

3. Genetic programming for job shop scheduling

   1) Scheduling heuristics for different types of job shop scheduling
   2) Genetic programming to learn scheduling heuristics
   3) Set up genetic programming as a hyper-heuristic approach for job shop scheduling

4. Surrogate-Assisted genetic programming for job shop scheduling

   1) Surrogate basic concepts
   2) Phenotype VS Genotype of genetic programming individuals
   3) K-nearest neighbour based surrogate for genetic programming
   4) Instance rotation based surrogate in genetic programming
   5) Simplified model based surrogate for genetic programming
   6) Collaborative multi-fidelity based surrogates for genetic programming

5. Genetic programming with feature selection for job shop scheduling

   1) Feature selection basic concepts
   2) Feature frequency-based feature selection
   3) Contribution-based feature selection

6. Multitask genetic programming for job shop scheduling

   1) Multitask basic concepts
   2) Multitask genetic programming based generative hyper-heuristics
   3) Surrogate-Assisted evolutionary multitask genetic programming
   4) Task related based multitask genetic programming
   5) Multitask multi-objective genetic programming

7. Future directions

   1) Limitations of existing studies in this field
   2) Potential solutions to handle the limitations

Organisers

Dr. Fangfang Zhang, Victoria Univeristy of Wellington
Prof. Mengjie Zhang, Victoria Univeristy of Wellington
Prof. Yi Mei, Victoria Universty of Wellington
Dr. Su Nguyen, La Trobe University