Ambiguity

This is a Haskell library to expose the "ambiguous random value generator". The mathematical and practical properties of this generator are available in our paper which you can find here: https://doi.org/10.1287/mnsc.1100.1307

We would be grateful if you would cite us in work that uses this ambiguity generator. We have provided the BibTeX infno here:

@article{AmbiguityInTheLabMS2011,
author = {Jack Stecher and Timothy Shields and John Dickhaut},
title = {Generating Ambiguity in the Laboratory},
journal = {Management Science},
volume = {57},
number = {4},
pages = {705-712},
year = {2011},
doi = {10.1287/mnsc.1100.1307},

URL = {
        https://doi.org/10.1287/mnsc.1100.1307
   
},
eprint = {
        https://doi.org/10.1287/mnsc.1100.1307
   
}

this info is also available in the ambiguity.bib file, if you prefer.

Trying out the Ambiguity Generator

There is a website where you can try out the ambiguity generator here:

https://ambiguity.chapman.edu

Installation

You can install the AmbiguityGenerator using the source code. You will need Haskell for the install. We recommend using stack which will be able to fetch the Haskell compiler for you.

After installing stack you should fetch the source code:

git clone https://github.com/HaskellAmbiguity/AmbiguityGenerator.git

Once you have the source code, enter the directory and build the applications using stack:

cd AmbiguityGenerator
stack build

If you want to install the applications locally you can also use:

stack install

By default this will install the packages in ~/.local/bin, so you will need to add that to your PATH.

If you just use stack build you can run the programs using stack exec, which we'll use in the following examples.

Histogram and Graph Generation

There's an application for generating graphs of realizations, and histograms of the draws under finite support. It can be run as follows:

stack exec histogram <runs> <samples> <range> <output-file>

• <runs> should be replaced with the number of simulations you wish to run.
• <samples> is the number of draws from the ambiguity generator.
• <range> is a number for the range of finite support values. E.g., 10 will be 0 to 9.
• <output-file> is the name of the output file. Something like graphs.svg will be good.

The code for this is in the app/ directory.

Generating Draws

If you just want to generate a bunch of CSV files of ambiguous data, you can use the draws program:

stack exec draws <runs> <samples> <file-base>

• <runs> should be replaced with the number of simulations you wish to run.
• <samples> is the number of draws from the ambiguity generator.
• <file-base> is the string at the start of the output file.

This will generate a number of CSV files in the current directory of the forms:

• <fileBase><samples>-bits-<run>.csv: coin flips.
• <fileBase><samples>-digits-<run>.csv: values from 0 to 9.
• <fileBase><samples>-ambiguous-<run>.csv: raw realizations.

The code for this program is in the draws/ directory.

Web Server

There is a web application which allows users to download CSV files of certain draws. Additionally this server hosts an API that can be used to get output from the ambiguity generator in a JSON format.

stack exec ambiguity-server 8080

Will run the web server on port 8080. You can then view the website by going to: http://localhost:8080

The code for the web server is in the web/ directory.

Library

The ambiguity generator library is in src/.