Conference Agenda

This presentation delves into the transformative process of turning rapid prototypes into market-ready operations research (OR) applications, drawing from a variety of real-world projects. We focus on the methodical transition of prototypes to fully developed solutions, addressing recurring challenges and strategic solutions along the way.

We will explore two main areas: implementing effective OR solutions that meet dynamic market needs, and extracting insights crucial for our product development. This approach shapes a toolkit that is robust and adaptable to evolving technologies.

A significant part of our discussion will highlight the importance of rapid prototyping. This agile phase fosters informed discussions with clients, the end-users, guiding the refinement of prototypes to better meet their needs and expectations.

Furthermore, transitioning from a prototype to a mature OR application is a comprehensive development process. It includes enhancing user interfaces, optimizing deployment strategies (like GUI and cloud computing), and ensuring superior computational performance.

By sharing our experiences and best practices, this talk aims to provide participants with strategies to overcome common obstacles in OR project development. Attendees will gain a deeper understanding of how to effectively move from conceptual prototypes to advanced, market-ready applications, aligning with user needs and achieving operational efficiencies.

Often Operations Research applications suffer from not being able to quickly show the business impact in an integrated manner. This reduces the number of deployed applications. In this talk I will demonstrate how to reduce the development time, from model implementation, scenario analysis and closing the feedback loop with FICO Xpress Insight.

At the beginning we showcase the seamless deployment of a Python optimization model and transforming it into a fully operational application within Xpress Insight. We demonstrate the latest enhancements in our IDE Xpress Workbench boosting the Python model development experience, such as autocompletion and enhanced debugging views for various data types. Beyond local debugging, we will spotlight remote debugging capabilities, which allows debugging of scenarios while being executed in Xpress Insight on a remote server.

The Xpress Mosel language is used to implement large optimization and analytics projects. Many such projects are deployed as Xpress Insight apps, either on individual software installations or in a cloud-based environment.

To develop, build, deploy, and maintain such a large code base a team of OR specialists typically works in collaboration with other experts (UI specialists, problem domain experts). The availability of a suitable toolset for building, analyzing, and testing the code in a largely automated way is a must nowadays. The Mosel distribution comes with a set of open-source tools supporting these development tasks.

In this talk we review the roles and usage of the different tools to support the operation of enterprise-grade OR applications. We start with the build process (including the generation of online documentation), we then move on to the topics of testing and code analysis (profiling and coverage) and finally debugging. We consider in particular the case of Insight apps.