Effective 10/13/2023, we’re pleased to share that the Microsoft Quantum Inspired Optimization (QIO) solver is now available on GitHub under the MIT license in collaboration with KPMG. We’re excited for vibrant, community based QIO innovation to continue and generate progress and applications. Access here on GitHub.
Building a more environmentally sustainable future requires having a plan. A plan for how human society can optimally minimize its negative environmental impacts while maximizing the value it derives from Earth’s natural systems. It requires conservation planners and policymakers to weigh up ecological, and economic data points against the socio-economic and geo-political reality that they live in. Finding good answers to complicated spatial planning questions requires large amounts of computing and state-of-the-art computer science techniques. Using techniques inspired by quantum methods, the Azure Quantum team at Microsoft has partnered with The Nature Conservancy and the Marxan community to help modernize the most commonly used suite of tools in the conservation planning space: The Marxan planning engine.
Microsoft engineers have improved the Marxan software suite’s speed and usability and set the foundation for an Azure cloud-based Marxan platform. This will bring conservation planning capabilities to a wider group of policymakers, government, and academic users; and improve the time to a good solution, enabling real-time decisions in a far wider set of cases.
Marxan is the analytic engine behind major conservation projects and helps planners to make the best planning decisions; maximizing biodiversity and ecological protection. Today, Marxan is used by 1,300 organizations in over 100 countries around the world.
“We created Marxan to guide the resolution of Australia’s great forestry disputes in the early 1990s. Existing spatial planning tools didn’t deal with spatial relationships between locations or the variable costs to society of protecting different parcels of land. Marxan focused on delivering practical options to take to a negotiating table.”—Hugh Possingham, Chief Scientist, Queensland State Government and Marxan
Balancing conservation and economic sustainability across the globe
Marxan has been used to prioritize conservation actions on every continent and ocean basin and is often used by environmental non-governmental organizations (NGOs) to support government decision-making. For example, The Nature Conservancy used Marxan to plan a national conservation portfolio for Mongolia. The Mongolian Gobi Desert is part of the largest steppe ecosystem in the world, and the biodiversity and pastoral livelihoods of this area were threatened by rapid growth of mining and related infrastructure.
Image Credit: The Nature Conservancy
Image Credit: The Nature Conservancy
The Nature Conservancy worked with the Mongolian government, university researchers, NGOs, traditional pastoral communities, and industry to protect and conserve the largest remaining temperate grassland on the planet and contribute to designing one of the most representative networks of protected areas for ecosystems in the world.
Putting landscape-level mitigation into practice in Mongolia, Heiner et al. 2019.
Original Image: Mike Heiner (The Nature Conservancy); Mongolia’s protected area portfolio. Color edited for publishing.
Some of Marxan’s other major conservation projects include:
- Expanding the national rainforest planning system in Guyana.
- Adding 41,000km2 of marine protected area in the Seychelles.
- Rezoning of the Great Barrier Reef Marine Park in Australia.
“Marxan is a highly flexible and adaptive tool helping solve real-world problems. It enables governments to integrate spatial planning into their protected area goals, it prioritizes landscapes for alternative energy in our ‘Development by Design’ work and establishes multiple-use zoning plans for the oceans that respect traditional uses while conserving biodiversity.”—Dr. Jennifer McGowan, Managing Director of Marxan and spatial planner with The Nature Conservancy’s Global Science team
Capabilities for solving optimization problems inspired by nature
There is a rich history of finding inspiration from nature to solve hard combinatorial optimization problems like those found at the heart of conservation planning. Annealing and tempering are ancient metalworking methods that have inspired a multitude of optimization techniques. Many optimization algorithms, such as simulated annealing, parallel tempering Monte Carlo, or genetic algorithms mimic natural processes. Moving into the quantum age, optimizers have been developed that make use of quantum mechanics to accelerate optimization and escape local minima in the cost function landscape through quantum tunneling. Emulating these quantum effects on classical computers has led to the development of new types of quantum solutions that run on classical hardware, called quantum-inspired optimization (QIO) algorithms. These algorithms allow researchers and developers to exploit some of the advantages of quantum computing approaches today on classical based hardware, providing a speedup over traditional approaches.
Azure Quantum puts these powerful optimization tools and more in the hands of researchers and developers. Azure Quantum is the world’s first full-stack public cloud ecosystem for quantum solutions. From building QIO solutions to solve practical problems today in Azure to tools and capabilities to write quantum algorithms running on quantum hardware, Azure Quantum delivers powerful tools to realize impact from quantum-inspired algorithms today and build toward more powerful quantum solutions in the future.
Applying what we learned when building Azure Quantum
As open source software, Marxan has been built largely through a grassroots effort, with volunteer investments in the code base, companion tools and plug-ins, and new applications have come from dedicated individuals in the community. However, an ongoing challenge limiting its reach is that Marxan is a command-line application that requires extensive training to be accessible to new users. This has resulted in huge untapped potential for collaboration.
The team behind Azure Quantum worked closely with The Nature Conservancy and the Marxan community to create a new modern version of Marxan that applies the learnings from the development of Azure Quantum and the Microsoft QIO offering. The latest version does not only deliver an improved algorithmic engine, but also enables easier contributions and future more accessible cloud-hosted offerings that leverage Marxan.
Harnessing the power of Azure Quantum Optimization:
Marxan has always used a heuristic algorithm like the ones present in the Microsoft QIO today. The expertise gained from crafting the optimization solver offering helped the Microsoft Quantum team significantly improve the Marxan code base.
Additionally, the engineering team applied their expertise from building the Azure Quantum service to ensure scalability in highly parallelized setups and on modern central processing units (CPUs) like those found in the Microsoft QIO offering.
Easier to contribute:
Microsoft is committed to Open Source. Thanks to the team’s unique engineering capability, Marxan now has a modernized code base that makes it easier for the community to contribute in the future.
Easier to use:
To help users get started with the command-line tool, the project team worked closely with power users in the Marxan community and improved error messaging.
Faster and more efficient:
This effort has helped improve Marxan’s computational speed and efficiency on a range of features. For sufficiently large datasets, that makes the difference between giving a conservation planner near-real-time results and running a process for several 10’s of minutes.
Feature-specific improvements:
| Functionality | V2 | V4 | Speedup Factor |
| Asymmetric Connectivity (Marxan) | 23mins | 1min | 23x |
| Preprocessing Species Penalty (Marxan) | 10mins 11s | 1min 36s | 6.4x |
| Preprocessing Species Penalty (Marzone) | 11mins | 12s | 55x |
All benchmarks performed on a 64 core Standard D64ds_v4 Azure VM
Scalability comparisons:
| Marxan Dataset | Repeats | Marxan v2 time | Marxan v4 time | Speedup factor |
| Dataset 1 | 1000 | 234s | 5s | 46.8x |
| Dataset 2 | 100 | 6000min | 98min | 61.2x |
| Dataset 3 | 100 | 1000min | 12min | 83.3x |
All benchmarks performed on a 64 core Standard D64ds_v4 Azure VM
| Marzone Dataset | Repeats | Marzone v2 | Marzone v4 | Speedup factor |
| Dataset 1 | 500 | 1231s | 13s | 94.7x |
| Dataset 2 | 500 | 2826s | 60s | 47.1x |
| Dataset 3 | 500 | 621s | 12s | 51.8x |
All benchmarks performed on a 64 core Standard D64ds_v4 Azure VM
A platform designed for today’s increasingly complex threats
The Nature Conservancy and Microsoft Quantum have partnered to deliver significantly improved analytical tools to the conservation planning community at a time when the world is facing increasingly complex threats to biodiversity and livelihoods.
This new version of Marxan powered by the expertise and knowledge behind Azure Quantum will soon enable more accessible, scalable, cloud-based conservation planning for larger and more integrated solutions, and a much broader set of users around the globe.
“Marxan is a great example of how large environmental data sets, hosted on the cloud alongside powerful computational resources, can enable sustainability practitioners to do conservation planning at a scale that was unimaginable on laptops.”—Dan Morris, Principal Scientist, Microsoft AI for Earth
Get started today
Leverage powerful quantum-inspired optimization capabilities and get started on your quantum journey with Azure Quantum.
Learn more about how the Azure Quantum team works with customers and partners to solve some of the world’s most complex computational challenges.
- If you want to try Marxan, you can download the new version from the Marxan Solutions website.
- Marxan has always been and will continue to be open source. You can find the Marxan code here.
We would like to encourage all users to join the Marxan community of practice and share your Marxan stories on Twitter with @Marxan_Planning hashtag #Marxan.