Grab your badge, some coffee and get ready to join us for two days packed with learnings and networking!

Welcome to The Gurobi Decision Intelligence Summit! Hear the latest Gurobi Optimization news and updates from CEO Duke Perrucci.

Optimization is all around us—calculating the quickest route on your mobile phone, ensuring timely package deliveries, and streamlining your airline and hotel bookings. Optimization technology is being used in all those situations because it “creates value” – but what exactly does that mean? What’s the hidden connection between a deeply mathematical puzzle-solving tool, and your bottom-line results? What is the interplay between complexity and value?

The optimization market is changing as the compute landscape shifts to emerging technologies. High-performance computing is on the rise as GenAI is pushing widespread GPU and TPU demand. Niche technologies like quantum and neuromorphic are capturing the imagination of the next generation. Classical processors continue to be a work horse for enabling these specialized systems while not losing their edge on critical load types. This new landscape means an inherent change for how decision sciences teams and systems will be incorporated into business. As companies can address larger and more complex challenges, redefining what is tractable, they must also be agile enough to adopt new practices and adapt to the changing landscape.

Every few years, a new computing technology emerges with the promise to transform the world. Some technologies like the personal computer, the internet and the smartphone became indispensable tools in our work and personal lives. Others, like the CD-ROM, had limited impact. Here, we consider three emerging technologies that may transform optimization: GPU computing, Generative AI and Quantum Computing. We will learn how each may potentially benefit optimization, and whether they are likely to become the next big thing or whether they may become a footnote in computing.

Gurobi can handle a number of nonlinear functions in optimization models. Historically, Gurobi first supported convex quadratic function in the objective and constraints. With Gurobi 9.0, Gurobi introduced a global solver for non-convex quadratic models and automatic piecewise-linear approximations of common arithmetic functions. Gurobi 11.0 comes with support for global optimization of models containing univariate nonlinear functions . In this talk, we will review those various features and present our current work for tackling nonlinearities better, including a preview of Gurobi 12.0.

Methane emissions are known to contribute significantly to climate change. Despite its shorter atmospheric lifespan than carbon dioxide, methane is a highly potent greenhouse gas (GHG), with a 20-year warming potential 80 times that of carbon dioxide. According to the EPA, about one-third of the current GHG-induced warming is attributed to anthropogenic methane emissions, with oil and natural gas operations identified as the foremost industrial source of methane emissions across the U.S.

Most oil and gas production sites across the country (roughly 80%) are low-production well sites, yielding, on average, 15 barrels of oil equivalent per day or less. Despite contributing to only 6% of overall oil and gas production across the nation, the combined methane emissions from these well sites constitute about 50% of the total methane emissions from all oil and gas production sites in the United States. In addition, several studies have found that abandoned and orphaned wells – which produce no hydrocarbons at all – can also release substantial quantities of methane into the atmosphere. Relative to their energy output, these low- and non-producing wells emit a disproportionate amount of methane, which makes them a particularly attractive target for environmental mitigation.

The U.S. Department of Energy’s (DOE) National Energy Technology Laboratory (NETL) is developing and releasing a portfolio of software tools to help reduce methane emissions from oil & gas operations. Specifically, NETL is making available a suite of free and open-source optimization-based decision-support tools that can help identify, characterize, prioritize, and mitigate methane emissions from oil & gas wells. These tools are designed to make recommendations about (1) which wells to target for “retirement” (known as permanent “plugging” in the industry), (2) how to make best use of precious well characterization and emissions quantification resources, and ultimately (3) how to design and execute efficient and impactful well plugging campaigns.

This talk will provide an example of one such optimization tool that NETL has developed to help guide emissions mitigation efforts across the oil & gas sector. It will highlight specifically how this tool addresses the many challenges that industry and state regulators face in this space, and why mathematical optimization – due to its distinctive attributes – is particularly suitable for meeting the technical and non-technical needs of decision-makers.

In the dynamic and competitive pulp and paper industry, integrated decision systems are crucial for companies to run efficient supply chain operations and establish a solid market position. These systems can assist decisions at several granularity levels reducing costs and improving overall performance with an integrated operational perspective. At Suzano S.A., the in-house data team created the Pulp Planning Portal in that sense, applying advanced analytical methods and tools, and leveraging numerical optimization models to assist operations across strategic, tactical, and operational levels. Exploring the flexibility ensured by mathematical models and the powerful solver performance of Gurobi, their tailor-made solutions are easily adapted to complex ever-changing business requirements.

At the strategic level, the models focus on long-term decisions such as capacity planning, resource allocation, production, product replacements, annual contract rules, and restricted demand planning on monthly buckets. The outcome of strategic models goes to the tactical level where mid-term decisions are taken on weekly buckets, such as vessel nomination, and inventory management. These results are unfolded into short-term operations planning on daily buckets with detailed production planning, transportation logistics, and vessel loading.

Based on experience implementing these systems, in this presentation, we aim to share how they can run in an integrated manner and how they help Suzano S.A. to solidify its position as a market leader in the pulp and paper industry. The path towards Suzano’s current technology position will be presented showcasing the main challenges and insight on both business and technical aspects.

Join us to explore the journey from conceptualization to implementation of these decision systems and discover the tangible benefits they bring to the industry.

This presentation examines the role of optimization tools in the renewable energy sector, focusing on how BrightNight’s PowerAlpha platform, powered by Gurobi’s optimization engine, transforms the design and operation of hybrid renewable power projects. As the energy industry targets greater sustainability and reliability, advanced decision-making tools are becoming increasingly essential.

PowerAlpha integrates advanced engineering with financial modeling to align fluctuating renewable energy production with grid demands. By leveraging Gurobi’s optimization engine, it simulates millions of design scenarios to find the most cost-effective configurations, ensuring projects achieve top-tier Levelized Cost of Energy (LCOE) and maximize returns.

Through case studies, the presentation will showcase how PowerAlpha optimizes solar and storage systems, delivering reliable power to data centers while achieving significant cost savings. PowerAlpha offers a comprehensive solution to the challenges of renewable energy project development, setting a new standard in the industry.

In this presentation, we explore the transformative potential of combining Causal AI, Automated Reasoning, and advanced Optimization to tackle complex business problems. We delve into how utilizing cutting-edge optimization techniques, integrated with Causal AI, drives improved decision-making, operational efficiency, and business growth. Attendees will gain insights into practical applications, illustrating how these technologies empower organizations to solve intricate challenges, improve profitability, and stay competitive in a dynamic market landscape.

Artificial Intelligence is rapidly transforming industry and manufacturing. At C3.ai, we provide artificial intelligence-based applications to enable this transformation. We will highlight a recent success providing an optimization application for sugar manufacturing that offers hourly AI-optimized recommendations to operators. The presentation will cover our dual approach of Machine Learning paired with Gurobi optimization for this customer solution.

The automotive industry faced unprecedented supply chain disruptions following the pandemic, impacting production and delivery across the sector. In this presentation, we will showcase how Toyota developed and leveraged optimization techniques to dynamically update its production plans to enhance supply chain resilience and adaptability. Additionally, we will also discuss how we further extended these capabilities to meet evolving customer demands as part of the broader Supply Chain and Fulfillment Transformation efforts. Attendees will gain insights into the practical application of optimization in navigating complex supply chain scenarios and adapting to a rapidly changing market landscape.

In the competitive landscape of manufacturing, the continuous improvement of processes is vital for maintaining an edge. This presentation delves into the intricacies of optimizing a drug manufacturing process through the lens of baking a cake. We will explore the trade-offs between different objective functions, the challenges of adoption, and the importance of a tangible and actionable solution. By employing advanced optimization models, we have demonstrated a 13% increase in throughput, in addition to massive time savings and error reduction. Join us as we showcase the transformative power of optimization in manufacturing, where strategic planning meets innovative technology to create a streamlined, cost-effective production process.

NVIDIA Grace™ is a groundbreaking Arm® CPU platform with uncompromising performance and efficiency. Discover how the NVIDIA Grace Family is revolutionizing the enterprise data center when paired with NVIDIA GPUs, such as the accelerated NVIDIA Grace Hopper Superchip, which delivers breakthroughs AI at scale, or as a standalone CPU for leading efficiency and scalability across the rest of the data center. Leverage a robust ecosystem of Arm solutions, OEMs, ISVs and system integrators to reduce your carbon footprint and transform your organization’s compute capabilities with a trusted, end-to-end solution.

Join us for dinner after a day of learnings and networking.

We’ll welcome you to Day 2 and will give you an overview of the tree tracks to choose from to ensure you get the most out of your day.

Our Journey Mapping sessions are for those just getting started with Optimization or are looking for advice on their path to greater adoption of Optimization within their organization. Schedule when registering for the Summit.

Development and Experts team members will meet with you to discuss your most pressing pain points.

Hear from the NFL’s Mike North, VP of Broadcast Planning, and Charlotte Carey, Director of Broadcast, on the latest learnings on how they helped to solve one of the hardest scheduling problems in existence.

In addition to solving LPs, MIPs and MIQCPs, Gurobi has many useful features that you may not be aware of. In this talk we review modeling features such as multiple objectives, multiple scenarios, solution pools and general constraints. We also present features to help analyzing infeasibility and tools designed to analyze and improve the performance of Gurobi on your models.

This session invites you to explore the complexities of decision-making through the Burrito Optimization Game, where you take on the role of a burrito truck owner, tasked with feeding hungry customers and ensuring tasty profits. Through gameplay and discussion, you’ll learn how some decisions can be made intuitively, while others—due to their complexity, tradeoffs, and time constraints—require optimization tools. After the game, we’ll tie these lessons to real-world problems, illustrating how similar challenges arise in areas like warehouse placement and oil well location.

This session will introduce you to the importance and fundamentals of Mathematical Optimization (MO). You’ll learn how MO differs from other methods in AI, how it aids in making complex decisions by translating predictions into actionable solutions, and why it’s so valuable to add this expertise to your professional toolkit. You’ll see how to identify the essential building blocks of optimization models in business problems: decision variables, constraints, and objective functions. This session will also feature a hands-in example, guiding you through running your ‘Hello World!’ optimization model.

Finding an optimal solution can be quite challenging for some models. In this session, we discuss potential reasons for this and how to deal with it. We also show a few techniques that help determine why some model is infeasible. Additionally, we give insights into the geometry of MIP to better understand the difficulty of models. Heuristics are essential to finding reasonable solutions quickly. We discuss heuristics included in the solver and other general concepts.

You already have an in-depth knowledge of your business challenges, likely supported by code and conditional statements that define the business rules and established specifications. Additionally, you have regression and machine learning models that provide predictions and address more intricate aspects of your business. In this session, we will guide you through converting these elements into the components of a robust optimization model. We will illustrate this by extending the example introduced in the ‘Optimization Crash Course’. By converting these existing representations into optimization frameworks, we’ll show you how to translate your predictive models into decision-making tools, embracing the principle that if you can model it, you can optimize it.

In the landscape of custom decision-support tools using commercial platforms, developers often face the challenge of navigating across multiple platforms tailored to specific analytics methodologies. As a result, optimization, ML, and simulation models tend to live in their preferred platform based on a natural division, which makes it challenging for users to interpret and reconcile the recommendations from these various models.

Gurobi’s available APIs in multiple programming languages provide a great opportunity to integrate optimization solutions in a variety of modern data warehouses and cloud architectures. In this talk, we demonstrate the benefits of Databricks through a network optimization example. Using this approach, we can streamline forecasting, optimization, and scenario analysis in a unifying platform, which can help foster better collaboration across teams with various analytics capabilities, and ultimately accelerates time-to-value in complex decision-making.

Imagine a project where issues are identified early, users are engaged and excited, and valuable results are delivered rapidly. In this talk, we’ll explore how adopting a customized agile methodology, specifically tailored for optimization models, can revolutionize their development compared to the traditional waterfall approach.

Agile for optimization involves iterative, modular development, frequent validation, and continuous integration to manage the unique complexities of optimization problems. This approach allows for faster delivery of value, enabling teams to see results and make improvements quickly. Enhanced collaboration and continuous feedback loops foster a user-centric approach, ensuring the model meets real-world needs and gains user buy-in.

By addressing potential issues early, agile reduces risk and improves overall model quality. Through engaging case studies, we’ll demonstrate how these benefits have been realized in actual projects, showcasing the success of agile for optimization in delivering robust and efficient solutions.

In today’s competitive landscape, optimizing business processes is crucial for success. This presentation explores the power of Supply Chain Optimization Systems (SCOS), leveraging mathematical optimization to improve planning and operations. We’ll delve into the key factors for identifying business challenges ripe for SCOS application.

Using a real-world case study of Vale, a leading Brazilian mining company, BITKA Analytics will showcase the development and deployment of a customized SCOS. This will demonstrate the tangible benefits achieved, overcoming challenges associated with implementing such a system.

Key takeaways:

• Identifying optimization opportunities within your business.
• Understanding the benefits and challenges of SCOS implementation.
• Real-world example of how SCOS empowers organizations.

Models with numerical issues can lead to undesirable results: slow performance, wrong answers or inconsistent behavior. When solving a model with numerical issues, tiny changes in the model or machine can make a big difference in the results. In this session, you will learn about Gurobi’s guidelines on numerical issues, how to identify them, how they impact your solutions and, most importantly, how to avoid them.

See the powerful potential and significant risks of using generative AI for modeling. While generative AI can help you get started modeling and coding quickly, it often requires careful oversight. We will walk you through a couple of examples of optimization modeling using Generative AI, highlighting key issues that can arise. The session will also cover quality assurance and best practices to ensure reliable and accurate models.

What is the business value of optimized decisions that don’t touch the real world? The path to delivering useful solutions in operational environments is well trod, but not always smooth. While decision optimization technology plays a critical role in driving cost savings in industries worldwide, there are tremendous — and often overlooked — gains to be had in streamlining the infrastructure, tooling, and collaboration workflows that increase the efficiency of operationalizing said technology. Similar to how the DevOps movement ignited a transformation in software development tools and practices, DecisionOps promises to not only simplify the process of shipping optimization models to production, but also scaling and accelerating model development with confidence and buy-in from stakeholders to ultimately derive more value out of optimization investments.

In this session, Gurobi Partners AMPL, Nextmv and Ormae, and Frontline Systems will host roundtables where you can bring your questions to get insights and recommendations for your toughest problems.

AMPL Roundtable topic: Streamline & Conquer: How Optimization Boosts Efficiency

In today’s competitive landscape, streamlining operations and maximizing resource efficiency are necessities. This round table, hosted by AMPL Optimization, explores how optimization transforms performance by finding the best solutions to complex problems. Whether it’s cutting costs, improving quality, or enhancing services, optimization empowers data-driven decisions for short- and long-term gains. Join us for a dynamic discussion to share challenges, insights, and explore solutions tailored to your operational needs.

Facilitator: Christian Valente, Senior Software Engineer, AMPL

Nextmv Roundtable topic: Supercharge your growth: How optimization drives top-line revenue

When optimization technology works well it feels magical. But it is not magic. From the modeling framework to the solver, infrastructure, workflows, and people contributing, decision optimization is team sport that involves orchestrating a symphony of moving parts. This means there are opportunities to optimize within optimization technology stacks and teams in order to drive top-line revenue. This roundtable will explore what those opportunities are through different lenses: technology choices, algorithm team dynamics and productivity, solution and results analysis, business stakeholder buy-in, collaboration across disciplines, and more. Bring your questions, observations, and stories of what has or had not worked, and walk away with tangible insights into how to improve your approach going forward.

Facilitators: Carolyn Mooney and Ryan O’Neil from Nextmv

ORMAE Topic: Mitigate Risk, Maximize Results: The Power of Optimization

Many organizations operate in multiple-geographies, and for them, a disruption in one corner of supply-chain can sometime unexpectedly turn into a major risk. For example, their manufacturing may be concentrated at a single location or probably they are dependent upon a single supplier for a critical raw material. In these situations, modelling capabilities of optimization can help them to identify what is the best way of demand fulfilment when the weakest link in their supply chain breaks. Also, modelling multiple scenarios can help organizations identify their weakest links.

Facilitator: Dr. Amit Garg, Founder and CEO, ORMAE

Frontline Systems Topic: Unleashing Cost Savings Through Business Optimization

This Roundtable will focus on applications of optimization that yield measurable cost savings – an outcome that businesses always want to achieve, but that often becomes urgent in case of a recession or industry slowdown.

Besides identifying common applications, found across industries, where cost savings can be realized via optimization, we’ll discuss the kinds of companies and people who are good prospects for an optimization solution. Wherever possible, we’ll aim for concrete use cases and demonstrations, not just abstract ideas, for how we can help customers achieve their goals.

We’ll start with the view that optimization provides better ways to allocate scarce resources to specific uses to reach business goals, subject to a range of constraints. The resources may be money, raw materials, inventory, warehouse space, production time and space, vehicle and equipment time, and often people time. Businesses seek to apply resources to achieve goals – for example meeting demand from customers – doing so at lowest cost. Optimization “shines” when there are far too many possible combinations of resource assignments to “figure out manually”, so an automated method is needed.

Some classic examples we can discuss in this Roundtable include:

• The “transportation problem” (shipping materials or products between facilities), which arises in many industries, where an optimization model can find a set of resource assignments (quantities of goods on routes) that minimizes shipping costs, or shipping and inventory holding costs.
• The “knapsack problem” (loading trucks, ships or containers with items of various sizes) is another problem that arises in many industries and settings. An optimization model can find the most efficient way to load each vehicle, meeting shipment goals and other constraints with the fewest (and hence least costly) number of trucks, ships or containers.
• The “blending problem” (selecting the right amounts of inputs – from crude oil or chemicals to food grains – to yield a mix that meets quantified quality goals at the least possible cost) is a third problem that arises in many industries – from refining oil to feeding farm animals to even blending stocks and bonds in an investment portfolio.

But not all businesses are equally good candidates for a cost-saving optimization model. Large businesses can very often benefit from a small percentage cost savings, because that percentage is applied to millions or even billions of dollars of cost. But for many small businesses where costs are measured in thousands of dollars, the savings achievable from optimization may not justify the cost of software and the user / analyst time and effort to build and test the model. Businesses also vary in their ability to pull together the data needed to drive the optimization model, on a timely basis.

This Roundtable will also seek to address that last problem: How to lower the user / analyst time and effort required to build and test an optimization model, thereby enlarging the available market. We can discuss pre-packaged application solutions and pre-designed models, the use of spreadsheets, high-level modeling languages, and high-level programming language and Internet APIs to lower time and cost. We can also discuss user training programs, the role of higher education, software “wizards”, and modern generative AI tools to speed and simplify the task of building and testing optimization models.

Facilitator: Daniel Fylstra, Founder and CEO, Frontline Systems

In this session, we will share insights and lessons learned from helping Gurobi customers from a wide range of industries adjust their optimization models to improve solver performance and numerical behavior. We will look at the challenges that we see most often in LP, MIP and MINLP models, and discuss our approach and typical recommendations to help address them. We will also consider some well-known modeling “rules of thumb” and discuss how applicable they are in 2024.

An invited panel of experts will discuss strategies and experiences getting buy-in for optimization projects. Analytics professionals and projects maximize their impact when they get the right support. Panelists will discuss strategies for communicating the impact and value of optimization projects, generating excitement, and getting buy-in from leadership and business stakeholders.