ARMLIB 0.1

Description

This library is a collection of spatially constrained forest harvest scheduling problems. The problems are instances of the Area Restriction Model (ARM) and are mostly based on datasets from the Forest Management Optimization Site (FMOS).

The objective of the ARMLIB collection is to serve as a repository of ARM instances used in the literature. In addition, we hope the collection will stimulate research in the area by presenting challenging instances in similar way to what MIPLIB has done for the mixed integer programming comunity.

The current maintainers of the collection are Marcos Goycoolea and Juan Pablo Vielma. If you have any questions or suggestions or if you would like to contribute an instance please send a message to

Version 0.1

Version 0.1 is the first version of ARMLIB. Its main objective is to introduce the file format of the collection, provide some illustrative instances and encourage the comunity to share their instances in this format.

File format

ARMLIB distinguishes between DataSets and Instances. DataSets include the characteristics of a forest map such as stand attributes and adjacency relations between stands. DataSets of ARMLIB are usually directly obtained from an FMOS datasets by maybe fixing some assumptions about stand attributes or adjacency definitions. DataSets are composed of a directory or folder containing several files describing the DataSet.

Instances describe exactly the ARM problem that should be solved. An Instance is based in a particular DataSet and describes the objective and constraints of the ARM problem to be solved. An instance is composed of a file with a .arm extension. ARMLIB DataSets and Instances formats are described in detail in the following pdf file.

ARMLIB 0.1 File Format

A simple example of the file format and other auxiliary material such as scripts, tutorials and MPS files of some mixed integer programming formulations of the instances can be found here.

DataSets and Instances

Instance difficulty