Evolutionary Optimization Algorithms: Fundamentals
Evolutionary algorithms (EAs) are approaches to artificial intelligence that are motivated by optimization processes that we observe in nature, such as natural selection, species migration, bird swarms, human culture, and ant colonies. This two-day course provides a clear explanation of the basic principles of EAs. The course covers the theory, history, mathematics, and application of EAs to engineering optimization problems. Featured techniques include genetic algorithms, evolutionary programming, evolution strategies, ant colony optimization, particle swarm optimization, differential evolution, biogeography-based optimization, and many others. Matlab-based examples are used during the course to illustrate the algorithms. This application-oriented course helps the student obtain a clear, but theoretically rigorous, understanding of EAs. The course also discusses the similarities and differences between various EAs. This course provides an ideal EA introduction to engineering and computer science professionals. Each student will receive a copy of the text Evolutionary Optimization Algorithms written by the course instructor, Dan Simon, in addition to a complete set of lecture notes and Matlab code.
What you will learn:
- The difference between evolutionary algorithms (EAs), computer intelligence, population based algorithms, biologically-inspired algorithms, and swarm intelligence.
- The four fundamental EAs.
- Design and program an EA for my problem.
- Some of the important tuning parameters in EAs.
- Latest EA techniques.
- Similarities and differences between various EA techniques.
- The no free lunch theorem and what are its implications for EAs.
- Perform a statistically rigorous comparison between the performance of different EAs.
- Introduction. Terminology. Unconstrained optimization. Constrained optimization. Multi-objective optimization. Multimodal optimization. Combinatorial optimization. Hill climbing algorithms.
- Genetic Algorithms. History. The binary GA. The continuous GA. Matlab examples.
- Performance Testing. Benchmarks. The no free lunch theorem. Overstatements based on simulation results. Random numbers. T tests. F tests.
- Evolutionary Programming. Continuous EP. Finite state machines. Discrete EP. The prisoner’s dilemma. The artificial ant problem.
- Evolution Strategies. The (1+1)-ES. The 1/5 rule. The (mu+1)-ES. The (mu+lambda)-ES. The (mu,lambda)-ES. Self-adaptive ES.
- Evolutionary Algorithm Variations. Initialization. Convergence criteria. Problem representation. Elitism. Steady-state vs. generational EAs. Population diversity. Selection options. Recombination options. Mutation.
- Ant Colony Optimization. Pheromone models. The ant system. Continuous optimization. Other ACO models.
- Particle Swarm Optimization. The basic PSO algorithm. Velocity limiting. Inertia weighting. Constriction coefficients. Global velocity updates. The fully informed PSO algorithm. Learning from mistakes.
- Differential Evolution. The basic DE algorithm. DE variations. Discrete optimization. DE and GAs.
- Biogeography-Based Optimization. Biogeography in nature. The basic BBO algorithm. BBO migration curves. Blended migration. BBO variations. BBO and GAs.
- Other Evolutionary Algorithms. Genetic programming. Simulated annealing. Estimation of distribution algorithms. Cultural algorithms. Opposition-based learning. Tabu search. The artificial fish swarm algorithm. The group search optimizer. The shuffled frog leaping algorithm. The firefly algorithm. Bacterial foraging optimization. The artificial bee colony algorithm. The gravitational search algorithm. Harmony search. Teaching-learning-based optimization.
- Practical Advice. Software bugs. Randomness. The nonlinearity of EA tuning. Information in an EA population. Diversity. Problem-specific information.
REGISTRATION: There is no obligation or payment required to enter the Registration for an actively scheduled course. We understand that you may need approvals but please register as early as possible or contact us so we know of your interest in this course offering.
SCHEDULING: If this course is not on the current schedule of open enrollment courses and you are interested in attending this or another course as an open enrollment, please contact us at (410)956-8805 or email@example.com. Please indicate the course name, number of students who wish to participate. and a preferred time frame. ATI typically schedules open enrollment courses with a 3-5 month lead-time. To express your interest in an open enrollment course not on our current schedule, please email us at firstname.lastname@example.org.
Dan Simon has worked in industry, academia, and consulting since 1983. He has applied evolutionary algorithms (EAs) to problems such as missile tracking, prosthetic leg control, electrocardiogram diagnosis, robot control, aircraft engine diagnostics, electric power management and distribution, and automotive engine control. Dr. Simon is currently a professor in the Electrical and Computer Engineering Department at Cleveland State University in Cleveland, Ohio. He has written over 80 peer-reviewed journal and conference papers, and has supervised over 20 graduate theses and dissertations. Dr. Simon is the author of the textbooks Optimal State Estimation (John Wiley & Sons, 2006) and Evolutionary Optimization Algorithms (John Wiley & Sons, 2013).
Contact this instructor (please mention course name in the subject line)