Maximize AI benefits with optimized decisions powered by CPLEX

By | 4 minute read | October 15, 2020

To better serve client needs, organizations must be able to access and analyze all types of data, no matter where it’s stored, whether data resides on premises, on multiple clouds, or scattered across both. IBM has developed a modern AI platform to help customers drive digital transformations throughout their organizations, giving them the right tools to apply data management and data science to their data strategies. Gartner defines an ideal Data Science and Machine Learning platform as “a cohesive software application that offers a mixture of basic building blocks essential both for creating many kinds of data science solutions and incorporating such solutions into business processes, surrounding infrastructure and products.”

IBM CPLEX is an integral part of IBM Cloud Pak for Data

IBM® brings this comprehensive philosophy to IBM Cloud® Pak for Data, a fully-integrated data and AI platform that helps businesses modernize, manage, govern and integrate their data so they can infuse AI throughout their organizations. A key differentiator of the platform is the IBM Decision Optimization portfolio, powered by the heritage of CPLEX® optimization engine and CPLEX Optimization Studio. The portfolio complements predictive analytics tools to help customers focus on developing data science platforms that better serve their clients. Using IBM CPLEX clients have realized significant benefits* through millions of dollars in cost savings*, up to 15% increase in productivity, less time to create financial portfolios and more. Furthermore, to help clients operationalize their models, IBM offers Watson Machine Learning, providing an easy-to-use, self-service approach to quickly and easily harness benefits of powerful CPLEX engines on the cloud.

As customers combine predictive and prescriptive analytics models to drive business decisions, they face a number of challenges. Finding and using data for both kinds of models in a seamless manner is extremely difficult. Predictive and prescriptive analytics are still fairly separate in terms of implementation, and one of the primary reasons is that data science and Operations Research teams are often physically and functionally separated in customer organizations. Data must be transported across different systems from varying vendors, to use insights gleaned from one method to use in another. By combining Decision Optimization technology in IBM Watson® Studio on Cloud Pak for Data, data scientists and operations research specialists can use the same tools to manage and shape data for modeling. Using the optimization and predictive model on the same platform helps facilitate outputs at a faster rate to help you make better, more insightful business decisions.

Overcome Data Science skills shortages with IBM Decision Optimization

A shortage of Data Science skills continues to plague enterprises, and in the Operations Research space this shortage is even more prevalent. In order to counter the skills gap, IBM is investing in technologies that enable higher productivity for skilled Operations Research scientists, as well as data scientists who may be new to the Operations Research space. Through a natural language based modeling assistant, customers can now easily create optimization models without having to write code. In addition, they can perform what-if analysis to understand the trade-offs and create and explore scenarios in the solution space. These tools drastically improve the time to value in this complex space facing a shortage of skills.

Constraints in real business environments are understood best by people on the frontlines — the operations executives and associates in lines of business. They must control optimizing and consuming business decisions provided by the AI platform. IBM Decision Optimization Center helps customers build and serve end-user applications, powered by the CPLEX optimization engine.

IBM Research investing to radically simplify data-driven decision making

IBM recognizes the need to automate and thoroughly simplify the arduous transition from data to decision making. IBM Research is actively developing solutions that automate the creation of decision optimization models for select data-driven optimization patterns that unite predictive and prescriptive modeling techniques. These solutions are designed to automatically combine best of breed automated machine learning models (regression and forecasting) for function approximation over input training data with matching mathematical programming models that are automatically formulated and solved. A related research effort, one that focuses on developing techniques that optimize machine learning loss functions, appears as a spotlight in NeurIPS 2020.  A complementary direction of research also underway brings best of breed optimization techniques for automated machine learning, included in a paper to be delivered at NeurIPS 2020, where we leverage scalable mixed-integer linear programming techniques to learn optimal multivariate decision trees.

Beyond the core technology stack, we are investing further to help customers derive the most from the AI platform with support from prescriptive analytics. Future integrations with business analytics tools will help users push data directly to Watson Studio and consume the output of optimization models within those tools. The IBM Data Science and AI Elite team continues to ramp up optimization skills to help customers operationalize their AI projects and drive business decisions.

IBM is committed to guide our Business Partners and customers on their journey to AI. Prescriptive analytics — and specifically, optimization technologies — provide a unique capability that can help our customers get the best from the platform to drive positive business impact. We invite you to try out Decision Optimization capabilities through the classic CPLEX Optimization Studio or through Watson Studio on IBM Cloud® or IBM Cloud Pak for Data.


Leading bulk tanker transportation company