Natural Resources Solutions - IBM Research Brazil - overview
Globally, the Natural Resources industry is currently worth $7T and continues to grow with the growth of population and expansion of global economies. Many natural resources are critical to human survival while others are vital to our modern lifestyles and, as a result, demand – including that for raw commodities such as oil, gas, metals, minerals, fuels, agricultural products, wood, and many others – has been increasing in velocity and complexity as a growing and industrializing world population commands more manufacturing, consumer goods, capital goods, construction, and energy.
Considering the increasing relevance of the Natural Resources industry to the local and global economies, one of the pillars of our Lab is dedicated to the intelligent management of natural resources, from discovery and exploration to production and logistics. The Natural Resources Solutions group at IBM Research Brazil aims at advancing the ongoing transformation of the Natural Resources industry as it moves into a knowledge-driven era fueled by unprecedented data availability.
Our research agenda is focused on addressing the scientific and technical challenges of the Natural Resources industry by developing Systems of Insights that combine data-driven and physically-based analytics with mathematical optimization techniques. To provide the appropriate support for the development, deployment and management of those systems, we are also investigating flexible, resilient and integrated platforms based on cloud and high performance computing. The main areas that we concentrate our efforts are:
Data-driven Analytics: In this area we develop advanced data-driven techniques for critical problems in natural resources industries, with a focus on predictive modeling to aid decision-making. Our aim is to combine state-of-the-art techniques from database management, statistics and machine learning to design scalable predictive modeling solutions that ingest and integrate large amounts of data from a variety of types: geospatial, temporal, textual and graphical. Specific applications of interest include reservoir characterization, seismic interpretation and predictive asset management.
Physically-based Analytics: In this area, we invent new technologies for creating and exploiting models that represent the physical behavior of systems, combining computing and domain expertise in Natural Resources. Typically, these models involve specific mathematical techniques to support simulations, tuned to the needs of natural resources (e.g. oil reservoir modeling and simulation). Likewise, these models rely on specific computing capabilities, for example in High Performance Computing, Cloud Computing, and Scientific Visualization. Of particular importance is the ability to link the models to the increasingly large amounts of data collected on the natural systems, thereby combining approaches for Big Data processing and analytics, and physical modeling.
Mathematical Optimization under Uncertainty: In this area we develop both theoretical and applied research on mathematical optimization focused on modeling frameworks and solution algorithms for linear, nonlinear and integer programming. The correct characterization and incorporation of relevant sources of uncertainty into mathematical optimization models is of particular interest to our research agenda as this is arguably a fundamental aspect of real-world decision-making processes.
Solutions Engineering: In this area we investigate new conceptual models and technologies to support software-intensive systems in Natural Resources. The ultimate goal is to devise new models, methods and software technologies capable of enabling and speeding up the development, integration and replication of advanced solutions in discovery, exploration and production of natural resources. We are particularly interested in research topics related to Human-Computer Interaction, Programming Models and Tools, and Industry-specific Semantic and Data Modeling.
Industrial Cloud Technologies: In this area we perform research and develop technologies for industries by exploring Cloud Computing and HPC resources. In particular, we are interested in technologies for dynamic management of traditional (virtual machines) and specialized resources, such as physical machines, fast disks, high-speed networks, and GPUs, in hybrid cloud environments, comprising public and on-premise infrastructures.