Use of Clustering and Interpolation Techniques for the Time-Efficient Simulation of Complex Models within Optimization Tasks

Abstract

Several widely used model optimization techniques such as, for instance, genetic algorithms, exploit on intelligent test of different input variables configurations. Such variables are fed to an arbitrary model and their effect is evaluated in terms of the output variables, in order to identify their optimal values according to some predetermined criteria. Unfortunately some models concern real world phenomena which involve a high number of input and output variables, whose interactions are complex. Consequently the simulations can be so time consuming that their use within an optimization procedure is unaffordable. In order to overcome this criticality, reducing the simulation time required for running the model within the optimization task, a novel method based on the combination of clustering and interpolation techniques is proposed. This technique is based on the use of a set of prerun simulations of the original model, which are firstly used to cluster the input space and to assign to each cluster a suitable output value within the output space. Subsequently, in the simulation phase, an ad-hoc interpolation is performed in order to provide the final simulation results. The proposed method has been tested on a complex model of a blast furnace within an optimization problem and has obtained good results in terms of accuracy and time-efficiency of the simulation.

Reducing the energy consumption and CO2 emissions of energy intensive industries through decision support systems – An example of application to the steel industry

“…The description of such work is outside of the scope of this paper, but a detailed discussion can be found in [39].…”

“…Many complex problems which represent real world operations, where no direct analytical form of the relationship between process inputs/outputs is available, make use of advanced optimisation techniques such as GAs, PSO or ACO [36][37][38][39][40]. All these techniques exploit several iterative runs of the model to test the effect of changes in a set of input parameters on the output variables of the model in their search for an optimum.…”

Spatial Clustering Strategies for Hierarchical Multi-Scale Modelling of Metal Plasticity

“…This significantly reduces the number of updates of the plastic anisotropy and subsequently the overall simulation time. Similar approaches to reduce computational cost of engineering applications based on clustering are reported in [8,10,11].…”

“…So, it is possible that the statically constructed clusters, formed at the first update of texture and plastic anisotropy, are not representative for the spatial variation in plastic properties during the rest of the simulation. Figure 8b presents a comparison between the approximation error in the affected equivalent plastic strain fieldē γ given by (10), and the corresponding gain in time g t , given by (11). We see that varying the number of clusters up to 100 has no significant effect on the approximation error as well as on the gain in time.…”

Spatial clustering strategies for hierarchical multi-scale modelling of metal plasticity

“…The proposed method in [58] efficiently extracts microstructural information by employing data clustering algorithms. Similar approaches to reduce computational cost of engineering applications based on clustering are reported in [26,59,60] This implementation allows us to handle data with structure or connectivity, which is crucial for our application. Indeed, the connectivity defined by the FE mesh imposes the connectivity of the clusters.…”