To get started with AMPL, you will need to learn the syntax of the language. Below you will find a quick introduction to the AMPL syntax and entities. More in-depth teachings is found in the Learn and Build tab of this resource page.
To get started with AMPL, you will need to learn the syntax of the language. Below you will find a quick introduction to the AMPL syntax and entities. More in-depth teachings is found in the Learn and Build tab of this resource page.
Open source, interactive, visual applications of your model.
Open source, interactive, visual applications of your model.
Open source, interactive, visual applications of your model.
A hands-on masterclass teaching you how to build a powerful network optimization solver from scratch by first creating a production planning solver and progressively expand it into a full network optimization solution.
by Christina Radu and Filipe Brandao | Youtube Recorded Masterclass
A hands-on masterclass teaching you how to build a powerful network optimization solver from scratch by first creating a production planning solver and progressively expand it into a full network optimization solution.
Throughout the sessions, you’ll develop a Streamlit app and complete interactive exercises designed to reinforce your understanding of key concepts.
by Gyorgy Matyasfalvi | Jupyter Notebook Series on Google Colab
In this six-part series, you’ll start with a straightforward facility location problem, demonstrating how to utilize AMPL and the amplpy module within a Jupyter Notebook to find a solution, then progress through the series into a two-stage stochastic programming problem.
From the third notebook onwards, the focus will pivot from modeling to algorithm development in AMPL, exploring four distinct approaches to implementing the Benders decomposition. This sequence allows the illustration of how subproblems can be solved in parallel. Lastly, you’ll learn the AMPLS library, which facilitates exporting AMPL model instances to persistent solver representations. This approach enables an exceptionally efficient Benders decomposition implementation.
A hands-on masterclass teaching you how to build a powerful network optimization solver from scratch by first creating a production planning solver and progressively expand it into a full network optimization solution.
Python | Upskill
by Postek, Zocca, Gromicho, Kantor with Filipe Brandao for AMPL additions.
A complete introduction to mathematical optimization with AMPL: A Mathematical Programming Language written by the founders of AMPL Optimization
AMPL | Introduction
by Robert Fourer, David M. Gay, and Brian W. Kernighan
A complete introduction to mathematical optimization with AMPL: A Mathematical Programming Language written by the founders of AMPL Optimization
Energy | Upskill
by Anthony Papavasiliou
An analysis of optimization models that are routinely used in electricity market operations.
Introduction to Optimization
by Fabio Schoen
An eBook on the art of modeling optimization problems. This book was used in the Optimization and Data Science Management course, Master degree in Management Engineering at the University of Florence.
A hands-on masterclass teaching you how to build a powerful network optimization solver from scratch by first creating a production planning solver and progressively expand it into a full network optimization solution.
Hitachi Energy’s GridView platform is one of the most widely used power system planning and analysis tools in the energy industry. Utilities, grid planners, and market operators rely on GridView to simulate electricity markets, evaluate transmission investments, and optimize power system operations.
To support increasingly complex energy systems, including renewable integration, market constraints, and large-scale transmission networks, GridView integrates AMPL as its optimization modeling engine.
Today, GridView powered by AMPL is deployed by more than 30 energy companies, supporting hundreds of analysts and planners managing some of the most complex power grids in the world.
→ 30+ energy companies using GridView optimization models
→ Hundreds of power system analysts running optimization scenarios
→ Millions of variables solved in large-scale grid models
→ ~10 minute solve times for MILP models with tens of thousands of integer variables
With millions of customer accounts and a rapidly growing user base, Dropbox needed a better way to assign sales representatives to territories. Manual and spreadsheet-based methods led to uneven workloads, difficulty prioritizing high-value accounts, and limited scalability as the organization expanded.
Dropbox built an internal AMPL optimization model to automate territory design. Using a mixed-integer linear model with up to 10,000 binary assignment variables per region, the team generated optimized account-to-representative assignments within practical runtimes, even at large scale.
The model integrates directly with Salesforce data extraction, Python-based account scoring, automated result delivery for sales leadership.
→ Automated, scalable territory design
→ Balanced rep workloads
→ Priority accounts aligned with top-performing reps
→ Seamless integration into existing analytics and CRM systems
Zara operates in a high-velocity retail environment where inventory decisions must be made within hours of receiving updated sales data. Manual allocation processes struggled to balance store demand, warehouse inventory limits, rapid trend shifts, and the risk of overstock or markdowns, across thousands of items and stores worldwide.
Zara partnered with researchers from UCLA and MIT to implement an AMPL-based optimization model for inventory allocation. The model integrates demand forecasting based on historical sales, store-level inventory data, warehouse availability constraints, piecewise-linear relationships between inventory and expected sales.
AMPL runs separate optimizations per item, resulting in approximately 15,000 optimized allocation decisions each week.
→ Real-time, data-driven shipment decisions
→ Strict enforcement of inventory and feasibility constraints
→ Scalable optimization across thousands of stores and products
→ User-friendly deployment for a 60-person allocation team
Young’s Plant Farm ships diverse plant products, ranging from seedlings to large specimens, directly to major retailers using its own fleet. Manual packing led to wasted rack space, under-utilized trucks, longer packing times, and increased shipping costs. The problem resembled a complex three-dimensional packing puzzle, constrained by time-sensitive delivery requirements.
Young’s implemented an AMPL-based packing optimization model integrated into a VBA-enhanced spreadsheet workflow.
The model dynamically generates optimal rack-loading plans, processing datasets ranging from hundreds of thousands to tens of millions of variables, while delivering practical runtimes for daily operational use.
→ Maximized rack space utilization
→ Reduced truck requirements and shipping costs
→ Faster packing and loading times
→ Seamless deployment through a familiar spreadsheet interface
→ Scalable optimization for fluctuating order volumes
A hands-on masterclass teaching you how to build a powerful network optimization solver from scratch by first creating a production planning solver and progressively expand it into a full network optimization solution.
Used by national & regional grid operators, TSOs, and large utilities to ensure grid stability with real-time power flow optimization. Generic tools struggle with the scale and complexity of these nonlinear problems.
AMPL is built for mission-critical grid management, where failure is not an option.
Used by energy producers, ISOs, and trading firms to optimize bidding strategies and manage risk in volatile energy markets. Generic tools fail to handle multi-period, stochastic, and regulatory constraints at scale.
AMPL ensures precision in market participation, maximizing profits while minimizing operational costs.
Used by utilities and renewable energy firms to optimize storage, balance grid fluctuations, and integrate wind and solar power. Generic tools struggle with multi-objective optimization and intermittent energy supply.
AMPL enables seamless renewable integration, ensuring efficiency, profitability, and grid stability.
Used by asset managers, hedge funds, and institutional investors to balance risk and return across multi-asset portfolios. Traditional tools struggle with large-scale constraints, multi-period rebalancing, and complex risk factors.
AMPL powers precise portfolio construction, enabling firms to optimize allocations, minimize risk exposure, and adapt to market conditions in real time.
Used by quantitative analysts and fund managers to minimize deviations from benchmark indices. Traditional approaches often fail to account for transaction costs, liquidity constraints, and real-world trading restrictions at scale.
AMPL delivers high-precision tracking error optimization, ensuring portfolio performance remains tightly aligned with benchmarks while maximizing efficiency.
Used by commodities traders, hedge funds, and financial institutions to optimize trading strategies in volatile global markets. Generic tools struggle with multi-period, stochastic constraints, and regulatory compliance.
AMPL ensures precision in commodities trading, allowing firms to optimize pricing, hedge against volatility, and maximize arbitrage opportunities across different exchanges.
Used by shipping companies, airlines, and freight operators to optimize cargo movement across air, sea, rail, and road networks. Traditional tools struggle with routing constraints, fluctuating demand, and real-time disruptions.
AMPL powers precise logistics planning, enabling companies to optimize fleet utilization, minimize transit costs, and adapt dynamically to supply chain bottlenecks and delays.
Used by manufacturers, retailers, and distributors to balance inventory levels across multi-echelon supply chains. Conventional systems fall short when managing dynamic demand shifts, supplier constraints, and perishability factors.
AMPL enables real-time inventory adjustments, reducing stockouts and excess holding costs while ensuring just-in-time delivery to meet customer demand efficiently.
Used by industrial manufacturers and large-scale production facilities to optimize factory schedules and resource utilization. Legacy approaches struggle with capacity constraints, lead time variability, and energy efficiency considerations.
AMPL facilitates seamless production planning, ensuring optimal resource allocation, reducing downtime, and synchronizing supply chain components for higher efficiency and throughput.
Industry decision science applications must be adaptable and accessible for stakeholder buy-in and collaboration. Integrating AMPL for modeling, Nextmv for deployment, and Streamlit for visualization enables teams to accelerate real-world impact.
Filipe Brandão (AMPL Head of Development) and Nicole Misek (Nextmv VP of Engineering) discuss the benefits of this integration, share insights, and walk through a live demo using a facility location optimization example with stochastic models and multiple constraints.
Operations Research education builds strong OR foundations but industry expects OR graduates to go further. Employers look for practitioners who can frame complex, real-world problems, communicate trade-offs, and develop optimization models that are deployable, not just solvable.
Explore AMPL’s full integration stack from Python, to commercial solvers, data connectors, deployment and more.
