15h30 - 15h55
Surrogate based Efficient Global MINLP Optimization and its application
Real industrial studies often boils down to complex optimization problems involving mixed variables and time consuming simulators. To deal with these difficulties we propose the use of a Gaussian process regression surrogate with a suitable kernel able to capture simultaneously the output correlations with respect to continuous and categorical/discrete inputs. The surrogate is integrated in the Efficient Global Optimization method based on the maximization of the Expected Improvement criterion. This maximization is a Mixed Integer Non-Linear problem which we tackle with an adequate optimizer : the Mesh Adaptive Direct Search, integrated in the NOMAD library. We implemented the proposed strategy and validated its accuracy on toy problems and real field cases.
15h55 - 16h20
Observability of Power Systems
In a power system network, the currents in transmission lines and voltages at buses can be provided by Phasor Measurement Units (PMUs). An n-channel PMU measures the voltage of the bus it is placed in and the current in n adjacent lines. Furthermore, based on two fundamental electrical circuit laws, additional voltages and currents can be obtained. We aim to determine a location for the PMUs such that full observability of the network is guaranteed at minimum cost. To that end, we solve a bilevel integer programming model for which we present an efficient cutting plane approach.
16h20 - 16h45
Uncertainty based Multidisciplinary Design Optimisation of Conventional Space Transportation Systems
An investigation of methods for the multidisciplinary design optimisation of conventional space transportation systems under model and environmental/operational conditions is presented. Simplified engineering models, including propulsion, mass, launch/mission, are coupled to cost/operational models, and then integrated within a multidisciplinary optimisation framework, including different approaches and methods for the uncertainty treatment. The objective is to evaluate the usability and usefulness of the probabilistic approaches for uncertainty quantification and compare them to evidence based approaches, when included into the design optimisation phase.