2. The research activity is based on advanced optimization and machine learning techniques,
and belongs to the field of operational research, i.e., that branch of applied mathematics in
which complex problems or systems are analyzed through suitable mathematical tools to
guide the decision-making process. Given the cross-disciplinary nature of the applications
of operations research, collaborations with researchers with different backgrounds (for
example: economists, engineers, mathematicians, doctors), are frequent, with projects in
various areas. Some current and past lines of research are listed below.
Optimization (determination of optimal parameters of mathematical models of complex
systems)
· In the economic field: determination of optimal parameters in insurance contracts and
pension systems; optimal drug pricing; optimal choices of production and consumption.
· In the industrial field: call access control and congestion control in telecommunications
networks; optimization of physical properties of complex materials for the design of
mechanical filters.
Game theory (analysis of the interaction between agents with or in the absence of
cooperation): in the economic field, solution of mathematical models related to pollution
and deforestation through the theory of non-cooperative dynamic games; study of
centrality measures of nodes of network for the analysis of road transport networks
through the theory of cooperative games.
Machine learning (automatic learning from data) for environmental, medical, economic,
biomechanical applications: classification of natural hazards; medical image classification
and segmentation; study of soft skills and input/output analysis; analysis of human
movement data collected with motion capture techniques.
Research
activity
3. example of application of machine learning
techniques to the analysis of movement
example of a public transport network
example of control of a spatial robot
Images
4. Implementation (mainly through the MATLAB software) of the optimization and machine learning
algorithms resulting from the own research activity
Technologies and services
5. Considering the growing availability of datasets from companies and organizations, machine
learning is becoming a knowledge-extraction tool that will be applied across many sectors to
understand their hidden potential, detect patterns hidden in the data, and make possible both
the personalization of the services and the control, also in real time, of environments and
objects. In the context of the strategies for problem formulation and solving based on Machine
Learning, also optimization plays a relevant role, sharing many of the approaches expressed
above.
The research activities have applications in the automatic analysis and optimization of industrial
processes and complex systems. The following are some possible examples:
· Application of machine learning techniques to various phases of industrial processes (e.g.,
quality control, fault detection).
· Applications of operations research to the analysis of transport and telecommunications
networks (for example: routing problems, study of centrality measures of nodes on a graph,
traffic congestion analysis).
· Applications of optimization to logistics (for example: warehouse management, scheduling,
distribution).
By the way of example, techniques that can be used for such applications include:
unsupervised learning; supervised learning; semi-supervised learning; online learning; deep
learning; optimal control; reinforcement learning; cooperative game theory; non-cooperative
game theory; linear optimization; non-linear optimization; combinatorial optimization.
Applications
and
collaborations
6. Ufficio di Trasferimento Tecnologico della Scuola IMT Alti Studi di Lucca
Address: Piazza San Ponziano 6 – 55100 Lucca, LU
Web site: https://www.imtlucca.it/
E-mail: projects@imtlucca.it
Ufficio Regionale di Trasferimento Tecnologico
Address: Via Luigi Carlo Farini, 8 - 50121 Firenze, FI
E-mail: urtt.@regione.toscana.it
For more information
For more information