Uncertainty-Aware Planning and Scheduling in the Process Industries

Lead Research Organisation: Imperial College London
Department Name: Dept of Computing


Process planning and scheduling problems are becoming increasingly complex due to the expanding production and customer base around the globe. A decision maker is continuously faced with the challenge to optimise the production plans and reduce costs under uncertainty. The uncertainty can be attributed to factors including volatile customer demands, variations in the process performance, fluctuations in socio-economics around the locations of the production plants, etc. Another complicating issue is the time-scale at which the decisions have to be taken and implemented. Not being able to effectively take these issues into account can lead to increased costs, customer dissatisfaction, loss of competitive edge and eventually shutting down of the manufacturing bases. This project aims to develop planning and scheduling tools for optimal decision-making under uncertainty while taking into account the multiple time-scales. Each process planning and scheduling problem is unique and hence one modelling and model solution tool cannot address the peculiarities of each problem. A framework where uncertainties are classified into specific categories is the key to providing cutting-edge optimal solutions. So, a problem will have a number of uncertainties which will be classified based upon our proposed framework and then for each classification the appropriate solution methodology will be invoked. A hybrid uncertainty modelling and optimisation tool that exploits the synergies of the solution techniques for various classes of uncertainty will also be developed. The novel planning and scheduling tools developed in this project will be tested on real-life case studies from process industries from a wide variety of sectors including energy systems, agrochemicals, pharmaceuticals, consumer goods, oil & gas, and industrial gases. Optimal planning and scheduling solutions based upon personalised uncertainty will be obtained.


10 25 50
Description We investigate decision making under uncertainty, i.e. the ubiquitous process of selecting a course of action in order to solve a problem with incomplete information. For example, in industrial systems, meeting customer demands requires planning a production schedule which is robust to machine failures and unreliable customers. Mathematical optimisation provides powerful methods computing such schedules. We focus on re-optimization techniques which repair inefficiencies determined when unknown information is revealed. The benefit of this approach has been demonstrated in various settings, e.g. public and product transportation.

We show that re-optimization does not always help to overcome well-known theoretical computation obstacles in scheduling problems. However, if the initial planning is designed using lexicographic optimization, we show significant benefit. Lexicographic optimal solutions have nice structure in the spirit of the well-structured alphabetical order of dictionaries. Furthermore, in analogous way that a damaged dictionary remains a useful piece of knowledge, lexicographic optimization allows efficient re-optimisation under uncertainty. This work formalises, quantifies and proves mathematically these ideas. Furthermore, it develops a novel lexicographic optimisation method with a performance comparable to using powerful commercial software.
Exploiting sparsity with specialized Semidefinite Programming (SDP) relaxations had a huge impact on the application of SDP relaxations to realistic polynomial optimization problems. Indeed, when using the classical Lasserre hierarchy, it is only possible to solve problems with a few dimensions, but by exploiting the sparsity present in many applications, it is possible to solve problems with several hundred variables. In our research we argued that many applications have additional structure that can be exploited to similar effect.
In particular, many large scale polynomial optimization problems have their origins from the discretization of an infinite dimensional model.
The resulting finite dimensional model is sparse but has a large number of degrees of freedom. Optimization models that fit this class are boundary value problems, optimization with PDEs , optimal control, and Markov Decision Processes, among others.Despite the progress made in the last decade, it is still not possible to solve realistic instances of the models arising in these applications using sparse SDP relaxations. The main contribution of this work was to show how to take advantage of both the sparse and hierarchical structure present in many applications. Our theoretical results suggest that under appropriate conditions we should expect significant improvements in computational complexity. Our numerical results further support this claim, and we showed that a multigrid approach can improve the robustness and reduce the time required to solve large scale polynomial optimization problems.
Exploitation Route An interesting feature of our approach is its applicability to other problems arising in network communications and facility location. The polynomial optimisation problems and sum-of-squres application also have many applications in data science.
Sectors Chemicals,Digital/Communication/Information Technologies (including Software),Energy,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

URL http://wp.doc.ic.ac.uk/rmisener/
Description Our finding that lexicographic ordering helps recover better feasible solutions after an initial perturbation has been of significant interest to the Royal Mail Data Science Group. In collaboration with Royal Mail, we are now working to translate the research into practice.
First Year Of Impact 2017
Sector Digital/Communication/Information Technologies (including Software),Energy,Environment,Transport
Impact Types Societal,Economic,Policy & public services

Description BASF Research Project
Amount € 50,000 (EUR)
Organisation BASF 
Sector Private
Country Germany
Start 05/2017 
End 08/2017
Description Early Career Fellowship in Software development for novel engineering research
Amount £984,063 (GBP)
Funding ID EP/P016871/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 09/2017 
End 08/2022
Description Grand Challenges Research Fund
Amount £1,515,900 (GBP)
Funding ID EP/P029558/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 05/2017 
End 04/2020
Description Industrial CASE Account - Imperial College London 2017
Amount £1,499,328 (GBP)
Funding ID EP/R511961/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 09/2017 
End 09/2022
Description Industrial CASE studentship with Schlumberger
Amount £27,000 (GBP)
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 10/2017 
End 09/2020
Description BASF - Statistics and Machine Learning for Chemicals 
Organisation BASF
Country Germany 
Sector Private 
PI Contribution Using a problem relevant to BASF, my team investigates a large-scale industrially-relevant optimisation problem that includes machine-learning tree-based models in the objective function. Our methodology leverages decomposition structure in the optimisation problem. Our work directly impacts BASF's energy efficiency and is easily extended to other applications, e.g. materials design and finance, since tree-based models are well suited for modelling unknown nonlinear functions. Furthermore, our optimisation problem may broadly incorporate other machine-learning models, contributing to the design of a unifying framework where mathematical optimisation is integrated with machine-learning and data analytics for effective decision making. My team contributed the expertise and developed the methodology.
Collaborator Contribution BASF finds machine-learning tree-based models effective for modelling catalyst behaviour because closed-form mathematical expressions are not known for many chemical catalysis applications. Catalysts are essential for energy efficiency at BASF. Therefore, developing the best-performing catalysts requires optimising over their tree-based models. BASF proposed a large-scale industrially-relevant optimisation problem that contains machine-learning tree-based models in the objective function. The optimisation problem considers the tree-based models but also penalises solutions that are far from training data, i.e. it optimises the tree-based models closer to the experimental data. BASF also contributed £40k in initial seed funding to fund my PhD student associated with this project.
Impact Based on the outcomes of this project, BASF has now funded a full PhD scholarship (£270k, 2019 - 2023). This PhD student is associated with the EPSRC projects and partially trained by researchers associated with the EPSRC projects. PhD student Miten Mistry (paid by BASF) collaborated with PDRA Dimitris Letsios and PI Ruth Misener to develop a submission to the 2019 STEM for Britain competition. Miten's accomplishment is documented here: http://www.imperial.ac.uk/news/190330/department-computing-researchers-selected-present-research/ Our join paper with the BASF team has been conditionally accepted to INFORMS Journal on Computing (https://arxiv.org/abs/1803.00952).
Start Year 2017
Description Collaboration with the research group of Processor Liesbet Geris 
Organisation University of Liege
Country Belgium 
Sector Academic/University 
PI Contribution We provided the expertise in optimisation under uncertainty.
Collaborator Contribution The group of Professor Liesbet Geris provided the application and the computing resources. PhD student Mohammad Mehrian visited the group of Ruth Misener to collaborate on this project. PhD student Simon Olofsson (not funded by the EPSRC, but Simon's training leads him to collaborate on the project) also visited the group of Professor Geris.
Impact We have published two full-length journal papers with Professor Geris' group: one is in Biotechnology & Bioengineering (2018) and the other is in IEEE Transactions on Biomedical Engineering (2019). We have also published several conference papers together.
Start Year 2016
Description Royal Mail Data Science Group 
Organisation Royal Mail plc
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution The Royal Mail wants to minimise the number of vans required at each of their "Delivery Offices" or DOs. Royal Mail has approximately 1250 Delivery Offices and 43k vans, so reducing the number of vans is a big cost (and environmental) savings for the company. My research team has developed a methodology for responding quickly to uncertainty. In our work, we design a schedule the night before based on historical data. Then, when anything goes wrong on the day, our recovery strategy quickly adapts to the change. We show that optimising the night-before schedule in a good way, i.e. with lexicographic optimisation, helps recovery on the day itself. Moreover, our results make Royal Mail more comfortable that they can reduce the number of vans without penalty.
Collaborator Contribution Our collaborators in the Royal Mail Data Science Group have meet with us for ~5 full days over the course of the last 6 months. Our collaborators also have provided us with test sets to try out our ideas.
Impact Royal Mail has offered to fund an EPSRC CDT PhD studentship for further collaborations. Royal Mail has also offered to write a letter supporting one of the REF Impact Case studies. MSc student Natasha Page, who collaborated with Royal Mail and EPSRC-funded researchers, was honoured for excellence in her MSc thesis by the Operational Research Society. Natasha was awarded "Runner-Up" to the 2019 May Hicks Award (https://www.theorsociety.com/what-we-do/awards-medals-and-scholarships/may-hicks-award/). The associated Imperial News Article explicitly credits the EPSRC funding (https://www.imperial.ac.uk/news/191227/department-computing-msc-student-honoured-joint/). MEng student Suraj G, who also collaborated with Royal Mail and the EPSRC-funded researchers, was honoured for excellence in his MEng thesis by the Imperial Department of Computing as one of the best theses in his year. Suraj won the "NewVoice Media Prize for Computing". The initial work with Royal Mail was accepted by the conference COCOA 2019 (http://cocoaconference.org/program.html, https://link.springer.com/chapter/10.1007/978-3-030-36412-0_6). The initial conference paper incorporates the work of Royal Mail, the EPSRC-funded researchers, and MSc student Natasha Page. We have also submitted a full length journal version of the work (preprint: https://arxiv.org/abs/1912.06862). The full length version additionally incorporates the work of MEng student Suraj G.
Start Year 2018
Description Interview for a film about the Royal Academy of Engineering 
Form Of Engagement Activity A broadcast e.g. TV/radio/film/podcast (other than news/press)
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact I appeared in a Royal Academy of Engineering video discussing my research.
Year(s) Of Engagement Activity 2017
URL https://www.youtube.com/watch?v=2gYp1nm6D3M
Description Keynote at the 2018 European Symposium for Computer-Aided Process Engineering 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact I will be giving a keynote at the 2018 European Symposium for Computer-Aided Process Engineering. This is my user community, i.e., the community of engineers who would be best suited to take the contributions to mathematical optimisation and computer science that my team has made and translate them widely into practice. I will be using the keynote (in June 2018) to discuss the outcomes of two EPSRC grants. I already know that the community is keenly interested in our research, but I hope to really fan the flames and get everyone on board with trying our new mathematical approaches.
Year(s) Of Engagement Activity 2018
URL https://www.tugraz.at/events/escape28/scientific-program/keynote-speakers/
Description Seminar at Carnegie Mellon 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact 80 postgraduate students and their supervisors attended a seminar given by Ruth Misener, which led to discussions on the topic and interest in future research
Year(s) Of Engagement Activity 2019
URL https://twitter.com/crislopeslara/status/1102990243752542208
Description Voice of the Future 2016 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Policymakers/politicians
Results and Impact Attending the 2016 "Voice of the Future". The event gives the opportunity for younger representatives of a wide range of science and engineering organisations to ask questions of leading figures dealing with science in Parliament and Government, such as MPs, Ministers, and scientific advisers - with the event chaired by a member of the House of Commons Science and Technology Select Committee. The event has been a huge success in previous years - last March's Voice of the Future was broadcast on BBC Parliament.
Year(s) Of Engagement Activity 2016
URL http://www.rsb.org.uk/policy/policy-events/voice-of-the-future