خط مشی دسترسیدرباره ما
ثبت نامثبت نام
راهنماراهنما
فارسی
ورودورود
صفحه اصلیصفحه اصلی
جستجوی مدارک
تمام متن
منابع دیجیتالی
رکورد قبلیرکورد بعدی
Document Type : Latin Dissertation
Language of Document : English
Record Number : 152296
Doc. No : ET24088
Main Entry : William Donald Smart
Title Proper : Making Reinforcement Leaning Work on Real Robots
Note : This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract : We begin this dissertation with a discussion of learning on real, physical robots inchapter 2. We then go on to introduce the specific learning paradigm, reinforcementlearning, that we will be using in chapter 3. In this chapter, we also discuss theshortcomings of this paradigm, in light of our chosen domain. In the following twochapters, we propose solutions for these problems and introduce HEDGER, our mainreinforcement learning algorithm and J AQ L, our framework for using this algorithmfor learning on robots. Chapter 6 shows the effectiveness of the individual featuresof HEDGER and JAQL, and looks at how they effect overall performance. We thengo on to provide experimental results for JAQL on two simulated and two real robottasks in chapter 7. Finally, chapter 8 summarizes the contributions made by thisdissertation, discusses their relevance and suggests fruitful directions for further work..
Subject : Electericl tess
: برق
electronic file name : TL48287.pdf
Title and statement of responsibility and : Making Reinforcement Leaning Work on Real Robots [Thesis]
 
 
 
(در صورت عدم وضوح تصویر اینجا را کلیک نمایید)