This OPTEC Special Interest Group aims to join OPTEC researchers working on distributed and parallel optimization problems, which naturally appear in many network-based and large-scale applications. In network-based applications, a number of spatially distributed and interconnected processors aim to cooperate in finding a global solution to a distributed optimization problem that depends on data which are locally available at each processor. Large-scale applications, on the other hand, involve the solution of very-high-dimensional optimization problems and may benefit from dividing the computations over multiple parallel processors. In both cases, a major issue is to limit the data communication between different processors as much as possible, either with the aim of limiting the communication bandwidth and power (in particular in wireless-network-based applications) or because the communication link between processors is the typical bottleneck in terms of computation time.
The work of this OPTEC Special Interest Group will be concentrated on three important themes:
- the formulation of distributed and parallel optimization problems,
- algorithms for solving distributed and parallel optimization problems,
- applications of distributed and parallel optimization.
Links with Working Groups
WG1: large-scale systems and networks, distributed optimization and control
WG2: modelling problems with high dimensionality and large data sets
WG3: parameter estimation for large-scale and multi-scale models
WP4: large-scale shape and topology optimization and large multi-scale simulations on massively parallel computer platforms
WP5: large-scale matrix equations and inequalities
Coordinator: Toon van Waterschoot
Members: Rodrigo Moraes, Paschalis Tsiaflakis, Toon van Waterschoot, Attila Kozma, Quoc Tran Dinh, Alexander Bertrand, Marc Moonen, Moritz Diehl
