MINLPLib.jq


MINLPLib.jq is a JSONiq library of Mixed-Integer and Continuous Nonlinear Programming Instances in the native DGAL format. The models in this library are sourced from MINLPLib and have been automatically converted from their original format.

Prerequisites

MINLPLib.jq is designed to be run using Unity DGMS, an open-source JSONiq analytics run-time environment for building model-driven decision guidance applications.

Installation

You can install MINLPLib.jq as a local dependency using the Node.js package manager npm or yarn.

First, initialize a new package (if not already exists):

``bash
mkdir my-pkg && cd my-pkg
npm init -y
`

Next, from your package's base directory, install MINLPLib.jq:

`bash
npm install minlplib.jq
`

Running a model

To run a model, from your package's base directory, execute the following command (replacing with the name of the MINLPLib model you want to run):

`bash
dgms run node_modules/minlplib.jq/bin/.jq
`

Alternatively, if you want to modify or extend these models, you can import them into your own JSONiq query module. For example, the following JSONiq query (st_e29-couenne.jq) can be used to solve the st_e29 model using Couenne, an exact solver for nonconvex MINLP.

`xquery
jsoniq version "1.0";

import module namespace a = "http://dgms.io/modules/analytics";

import module namespace ns = "http://dgms.io/contrib/models/minlplib/st_e29";

let $input := {
"x1": a:variable({ name: "x1", bounds: [0.0, 0.997] }),
"x2": a:variable({ name: "x2", bounds: [0.0, 0.9985] }),
"x3": a:variable({ name: "x3", bounds: [0.0, 0.9988] }),
"b4": a:variable({ name: "b4", domain: "binary", bounds: [0.0, 1.0] }),
"b5": a:variable({ name: "b5", domain: "binary", bounds: [0.0, 1.0] }),
"b6": a:variable({ name: "b6", domain: "binary", bounds: [0.0, 1.0] }),
"b7": a:variable({ name: "b7", domain: "binary", bounds: [0.0, 1.0] }),
"b8": a:variable({ name: "b8", domain: "binary", bounds: [0.0, 1.0] }),
"b9": a:variable({ name: "b9", domain: "binary", bounds: [0.0, 1.0] }),
"b10": a:variable({ name: "b10", domain: "binary", bounds: [0.0, 1.0] }),
"b11": a:variable({ name: "b11", domain: "binary", bounds: [0.0, 1.0] })
} return {
"obj": a:minimize({
model: ns:st_e29#1,
input: $input,
objective: function($output) { $output."obj" },
options: { solver: "couenne" }
})
}

``