Research title: Probabilistic Modelling for Industrial Optimisation
Start date: October 2013
Many industrial problems consist of more than one inseparable problem creating the need for methods that can better model interdependencies. Probabilistic modelling and simulation is increasingly used in complex learning and optimisation problems. This research focuses on using probabilistic model based search algorithm to solve multi-component optimisation problems.
State of the art approaches include: Bayesian and Markov network-based Estimation of Distribution Algorithms, Covariance Matrix Adaptation Evolution Strategy, response surface modelling, sampling and MCMC simulation. We however focus on using variants of the Estimation of Distribution Algorithm.
Dr Olivier Regnier-Coudert
- Enhanced Genetic Algorithm for Multi-Mode Resource Constrained Project Scheduling Problem. In Proceedings of the Companion Publication of the 2015 on Genetic and Evolutionary Computation Conference (pp. 745-746). ACM. http://dl.acm.org/citation.cfm?id=2765535
- Regnier-Coudert, O., McCall, J., & Ayodele, M. (2013, July). Geometric-based sampling for permutation optimization. In Proceedings of the 15th annual conference on Genetic and evolutionary computation (pp. 399-406). ACM. http://dl.acm.org/citation.cfm?id=2463422