Artificial Intelligence based multi-objective optimisation for energy management in dynamic flexible manufacturing systems

Lead Research Organisation: University of Glasgow
Department Name: School of Engineering

Abstract

The main goal of this project is to address the multi-objective dynamic flexible job shop scheduling
problem for reducing energy consumption and its related costs. The project aims to develop a
system that employs composite dispatching rules that include reduction of energy consumption as
its main objective. It is intended that such a system could be implemented in a flexible production
system in which job scheduling occurs at random or unpredictable times. The proposed dispatching
rule would prioritise all the jobs waiting for processing on a machine in the manufacturing system
while taking into account different attributes of the job and the machine, as well as time.
The manufacturing industry currently faces the dual challenge of increasing energy prices, and
regulatory mandates intended to reduce carbon emissions, which can prove problematic for many
enterprises. It is becoming increasingly necessary to explore the potential of reducing energy
consumption of industrial manufacturing at a system level, which has so far been largely ignored. At
this level, operational research methods can be employed as an effective energy-saving approach. In
the future, the requirement on manufacturing system flexibility within the system will be increased
to realise mass customisation and personalisation. On-line decision making and optimisation
techniques to accommodate these uncertainties and to maintain robustness of the flexible
manufacturing system is becoming increasingly important within the background of industry 4.0.
The application of multi-objective algorithms shows a great deal of promise at addressing the issues
faced by the manufacturing industry. By producing fast and elitist multi-objective scheduling
algorithms, it will be possible to optimise a number of factors, such as cost, energy consumption and
lead time. These optimisations will go a long way in improving the current state of manufacturing,
both by reducing environmental footprint and by reducing financial overheads.
A great deal of research has been conducted into the development of elitist heuristics and
evolutionary algorithms, and their operation is well understood. Existing dynamic scheduling
algorithms will be extended to address the uncertainties within the manufacturing system as a
benchmark. By targeting the practical shortcomings of these algorithms, it will be possible to
develop a system which avoids these problems, most notably: the high computational complexity of
the sorting.
A large part of the project will focus on expanding the past research into multi-objective
optimization problems and evolutionary algorithms. By developing understanding of the function of
jobs within the manufacturing environment, and the implementation mathematical modelling of the
job shop scheduling process. A focus on fast computation and implementation will be the heart of
this research, in order to maximise the applicability of the scheduling algorithms to a modern
manufacturing environment. In the context of a manufacturing environment, swift computation is of
paramount importance.

This project can be seen as a continuation of the research conducted by Dr. Ying Liu into the field of
meta-heuristics and industrial job scheduling. As such, much of Dr. Liu's research will be examined,
with the final intention of expanding it into an industry application."

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509668/1 30/09/2016 29/09/2021
2125600 Studentship EP/N509668/1 30/09/2018 29/03/2022 Patrick Blake