Name: Navigating Decision Trade-offs: Scratching the Surface with Math Optimization
Start: 2022-04-20T15:00:00Z
End: 2022-04-20T15:00:20.000Z
Location: BrightTALK
Rating: 4

With Gurobi’s decision intelligence technology, you can make optimal business decisions in seconds. From workforce scheduling, portfolio management, and marketing optimization, to supply chain design, and everything in between, Gurobi identifies your optimal solution, out of trillions of possibilities.  

As the leader in decision intelligence, Gurobi delivers easy-to-integrate, full-featured software and best-in-class support, with an industry-leading 98% customer satisfaction rating. 

Learn tips on how to choose a mathematical programming solver to help you tackle your complex business problems.

Choosing a Math Programming Solver

As the de-facto standard in Python for data manipulation and analysis, pandas is deeply integrated into many analytics projects and tools. Gurobi’s Python API, gurobipy, does not directly interact with standard pandas data structures, so incorporating decision optimization into data science applications is not as straightforward as it could be. For better interoperability between the two libraries, what is needed is a higher-level syntax to build optimization models directly from suitably structured pandas data.

Our solution is gurobipy-pandas, a convenient wrapper library to connect pandas with gurobipy. It enables users to efficiently build mathematical optimization models from data stored in DataFrames and Series, and to extract solutions as pandas objects. This webinar will walk through basic concepts and complete modeling examples to demonstrate best practices for using this library.

Building Gurobi Models from Pandas Dataframes

Machine learning has become a prevalent tool to provide predictive models in many applications, in this webinar relationships between variables of an optimization model in Gurobi will be discussed.

In recent years, machine learning has become a prevalent tool to provide predictive models in many applications. In this talk, we are interested in using such predictors to model relationships between variables of an optimization model in Gurobi. For example, a regression model may predict the demand of certain products as a function of their prices and marketing budgets among other features. We are interested in being able to build optimization models that embed the regression so that the inputs of the regression are decision variables, and the predicted demand can be satisfied.

We propose a python package that aims at making it easy to insert regression models trained by popular frameworks (e.g., scikit-learn, Keras, PyTorch) into a Gurobi model. The regression model may be a linear or logistic regression, a neural network, or based on decision trees.

Using Trained Machine Learning Predictors in Gurobi

There’s no questioning the massive impact advanced analytics in AI and ML have had on our lives, and from a business perspective, these technologies have made available an unprecedented wealth of information for decision-making.  

This information has evolved our ability to understand what can happen in the future. What hasn’t evolved is our approach to making decisions.  Business decisions are more complex than ever, yet a critical element of intelligent decision-making is overlooked – mathematical optimization (MO) 

This talk will address several key topics on incorporating MO with AI and ML, including: 

Defining MO 
Describing how MO and ML can work together 
MO case studies and success stories 
Facts about MO (e.g. scalability, guaranteed optimality) 
Learning MO and getting started

The most critical and FORGOTTEN element in advanced analytics: ML, AI and MO

By 2030, oil & gas plays across the United States are expected to produce over 60 million barrels of
water every day. Most of this water is characterized by high Total Dissolved Solids (TDS) concentrations,
exceeding 300,000 mg/L in some plays. Even though many upstream operators strive to reuse produced
water for hydraulic fracturing, disposal continues to be the leading water management practice across
the United States. This talk will focus on the development, demonstration, and deployment of a novel
optimization framework for produced water management and beneficial reuse.

The framework is designed to identify cost-effective and environmentally sustainable produced water management, treatment, and reuse solutions. Specifically, the proposed platform will support decision-makers with the: 

(1) coordination of produced water deliveries, 
(2) buildout of the produced water infrastructure in
terms of pipelines and storage facilities, 
(3) the selection of effective treatment technologies, 
(4) the placement and sizing of treatment facilities, 
(5) the identification of beneficial water reuse options, and
(6) the distribution of treated produced water and/or concentrated brine for beneficial reuse. 

This talk will also discuss opportunities and challenges associated with the recovery of critical minerals (e.g., Lithium) from oil & gas produced water.

Project PARETO – DOE’s Produced Water Optimization Initiative

Machine learning and deep learning have become standard tools in a data scientist’s toolbox, applied to generate insights into large amounts of data. But organizations are realizing that these technologies can’t do it all, and when decision making needs to be fast, reliable, and defendable, mathematical optimization is perfect complimentary tool to machine learning. 

Predictive models help answer the question of “What is likely to happen?” but also answering “What actions should be taken?” will allow organizations to add depth and value to their analytics projects where others can’t.

Why and How to Partner Machine Learning with Math Optimization

Solving Complex Challenges with Gurobi:

Data science and analytics teams are doing more with data than ever before, but more is being expected from such teams to differentiate themselves from their competition. All too often these analytics teams are ending their projects too soon – failing to consider the decisions their analyses inform. That’s where Decision Analytics comes in, and for the most complex problems, Mathematical Optimization. 

Mathematical optimization and Gurobi are fantastic partners to machine learning and can help data scientists get the most out of their predictive models by working where they work – in popular programming languages like Python and R.

Key topics: 
What are Decision Analytics, really?
What are some examples and common approaches used in Decision Analytics?
What is mathematical optimization? What are the problems it can address?
What tools can I use?
How can I get started using optimization and Gurobi?
Who else uses optimization and Gurobi?

The Data Science Key to an Unpredictable Future

The world is changing. You must prepare for the massive data changes that are sweeping the world, in this presentation Jerry Yurchisin will explain how data practitioners are the key to change.

Master The Art of Decision Analytics For Data Science

In this latest Data Science Central webinar, find out why companies like Uber, Air France, SAP, and the NFL use Prescriptive Analytics for continuous innovation.

Leading Data Scientists constantly expand their skill sets to provide more value to their organizations. The best strategy for taking on new tools is to focus on the return on investment of different project types. Mathematical Optimization delivers extreme ROI for organizations that use it. Data Scientists who make the leap into Mathematical Optimization suddenly discover dramatic cost savings and new efficiencies to offer their management.

Prescriptive Analytics: The Data Scientist's Master Key to a Turbulent Future

As business decisions are becoming more and more complicated, Decision Analytics are becoming increasingly important and a way analytics teams can differentiate themselves from the competition. When these decisions become extremely complex and it seems the trade-offs are too numerous and intricate to work through, mathematical optimization lends itself as great tool for this job. 

Mathematical optimization is a versatile tool that is easy to start using and adds huge value for any data scientist by addressing the part of a problem that is usually overlooked – the decision.

Navigating Decision Trade-offs: Scratching the Surface with Math Optimization

Data Analysis

Math

Data Science

Optimize

Decision Making

Analytics

Data

Practicing business intelligence allows your company to transform raw data into sets of insights for targeted business growth. The business intelligence and analytics community on BrightTALK is made up of thousands of data scientists, database administrators, business analysts and other data professionals. Find relevant webinars and videos on business analytics, business intelligence, data analysis and more presented by recognized thought leaders. Join the conversation by participating in live webinars and round table discussions.

Business Intelligence and Analytics

As an IT professional, many of the problems you face are multifaceted, complex and don’t lend themselves to simple solutions. The information technology community features useful and free information technology resources. Join to browse thousands of videos and webinars on ITIL best practices, IT security strategy and more presented by leading CTOs, CIOs and other technology experts.

Navigating Decision Trade-offs: Scratching the Surface with Math Optimization

Presented by

About this talk

More from this channel