multiRL0.2.3 package

Reinforcement Learning Tools for Multi-Armed Bandit

estimate_2_SBI

Simulated-Based Inference (SBI)

estimate

Estimate Methods

algorithm

Algorithm Packages

behrule

Behavior Rules

colnames

Column Names

control

Control Algorithm Behavior

estimate_2_RNN

Estimation Method: Recurrent Neural Network (RNN)

multiRL-package

multiRL: Reinforcement Learning Tools for Multi-Armed Bandit

params

Model Parameters

plot.multiRL.replay

plot.multiRL.replay

process_4_output_r

multiRL.output

process_5_metric

multiRL.metric

data

Dataset Structure

engine_ABC

The Engine of Approximate Bayesian Computation (ABC)

engine_RNN

The Engine of Recurrent Neural Network (RNN)

estimate_0_ENV

Tool for Generating an Environment for Models

estimate_1_LBI

Likelihood-Based Inference (LBI)

estimate_1_MAP

Estimation Method: Maximum A Posteriori (MAP)

estimate_1_MLE

Estimation Method: Maximum Likelihood Estimation (MLE)

estimate_2_ABC

Estimation Method: Approximate Bayesian Computation (ABC)

estimation_methods

Estimate Methods

fit_p

Step 3: Optimizing parameters to fit real data

func_alpha

Function: Learning Rate

func_beta

Function: Soft-Max

func_delta

Function: Upper-Confidence-Bound

func_epsilon

Function: ϵ\epsilon–first, Greedy, Decreasing

func_gamma

Function: Utility Function

func_zeta

Function: Decay Rate

funcs

Core Functions

policy

Policy of Agent

priors

Density and Random Function

process_1_input

multiRL.input

process_2_behrule

multiRL.behrule

process_3_record

multiRL.record

process_4_output_cpp

multiRL.output

rcv_d

Step 2: Generating fake data for parameter and model recovery

rpl_e

Step 4: Replaying the experiment with optimal parameters

RSTD

Risk Sensitive Model

run_m

Step 1: Building reinforcement learning model

settings

Settings of Model

summary-multiRL.model-method

summary

system

Cognitive Processing System

TD

Temporal Differences Model

Utility

Utility Model

A flexible general-purpose toolbox for implementing Rescorla-Wagner models in multi-armed bandit tasks. As the successor and functional extension of the 'binaryRL' package, 'multiRL' modularizes the Markov Decision Process (MDP) into six core components. This framework enables users to construct custom models via intuitive if-else syntax and define latent learning rules for agents. For parameter estimation, it provides both likelihood-based inference (MLE and MAP) and simulation-based inference (ABC and RNN), with full support for parallel processing across subjects. The workflow is highly standardized, featuring four main functions that strictly follow the four-step protocol (and ten rules) proposed by Wilson & Collins (2019) <doi:10.7554/eLife.49547>. Beyond the three built-in models (TD, RSTD, and Utility), users can easily derive new variants by declaring which variables are treated as free parameters.

  • Maintainer: YuKi
  • License: GPL-3
  • Last published: 2026-01-26