liveBook · Manning

Chapter 3: Design Custom Environments for Reinforcement Learning Algorithms

What you will learn in chapter 3 of Applied Reinforcement Learning

This introduction explains why reinforcement learning agents need simulated environments for safe trial-and-error learning in business settings.

After this chapter you will be able to

What it builds on

Section by section

Introduction

This introduction explains why reinforcement learning agents need simulated environments for safe trial-and-error learning in business settings.

Key terms: trial and error, policy, simulated environment, environment, sandbox, state dynamics, constraints, feedback

3.1 Conceptual framework for designing business environment

This section presents a framework for scoping a business problem, designing a reinforcement learning environment, and testing its behavior before training.

Key terms: Business Abstract Layer, modularity, entities, assumptions, stakeholder analysis, as-is diagram, key performance indicators, KPIs, Simulation Design Layer, state space

3.2 Warehouse order picking environment

This section applies the environment-design framework to a warehouse robot that picks an order, returns to a depot, and is tested before training.

Key terms: Business Abstract Layer, order fulfillment, depot, KPI, state space, reset(), episode, step(), transition logic, next state

3.3 Perishable product dynamic pricing environment

This section designs and tests a reinforcement learning environment for pricing a perishable product over a limited selling window.

Key terms: perishability, limited selling window, price-sensitive demand, pricing policy, transition dynamics, demand model, urgency factor, random noise, DynamicPricingEnv, state space

3.4 Trailer loading and packing environment

This section designs and sanity-checks a 3D trailer-packing environment for reinforcement learning.

Key terms: 3D Bin Packing Problem, space utilization, packed volume, orientations, PackingEnv, observation, 3D grids, action space, reset, placement validation

3.5 Summary

Where this is used later

Concepts explained in this chapter

← Chapter 2: Markov Decision Process: Turning Problems intro Solvable Models Chapter 4: Perfect Knowledge, Optimal Policy: Dynamic Programming →

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.