Datasets for the paper "Mining Markov Network Surrogates for Value-Added Optimisation" presented at the Genetic and Evolutionary Computation Conference (GECCO) 2016

Abstract

Model coefficients for the models that are analysed in the paper, and predicted/true fitness values for the validation data. As described in the paper: Surrogate fitness functions are a popular technique for speeding up metaheuristics, replacing calls to a costly fitness function with calls to a cheap model. However, surrogates also represent an explicit model of the fitness function, which can be exploited beyond approximating the fitness of solutions. The paper proposed that mining surrogate fitness models can yield useful additional information on the problem to the decision maker, adding value to the optimisation process. An existing fitness model based on Markov networks was presented and applied to the optimisation of glazing on a building facade. Analysis of the model revealed how its parameters point towards the global optima of the problem after only part of the optimisation run, and revealed useful properties like the relative sensitivities of the problem variables.