Simulation Methodology - An Overview

Explore the fundamentals of simulation methodology in this comprehensive lecture from the Theory of Reinforcement Learning Boot Camp. Delve into key concepts such as numerical integration, multiple integrals, high-dimensional integrals, and the Monte Carlo method. Learn about the Central Limit Theorem, Monte Carlo integration, and quasi-random integration techniques. Examine output analysis methods, including fixed sample size and smooth functions of expectations. Investigate non-regular inference and subsampling approaches. Gain insights into the challenges faced in massively parallel computing environments. Presented by Peter Glynn from Stanford University, this talk provides a thorough overview of simulation methodology, equipping learners with essential knowledge for advanced computational techniques in reinforcement learning and related fields.

Introduction
Outline
Terminology
Numerical Integration
Multiple Integrals
Highdimensional Integrals
Monica Law
Monte Carlo Method
Central Limit Theorem
Monte Carlo Integration
Quasirandom Integration
Output Analysis
Fixed Sample Size
Smooth Functions of Expectations
Nonregular Inference
Subsampling
Challenges in massively parallel computing