Delivering better and cheaper school bus services
Submitting Institution
University of SouthamptonUnit of Assessment
Business and Management StudiesSummary Impact Type
TechnologicalResearch Subject Area(s)
Mathematical Sciences: Applied Mathematics, Numerical and Computational Mathematics
Information and Computing Sciences: Artificial Intelligence and Image Processing
Summary of the impact
Research on designing mathematical methods for optimisation carried out
at the University of Southampton has been fundamental to the development
of software solutions for transportation problems and has directly led to
the growth and commercial success of the niche software company, Logical
Transport. Additional beneficiaries are local councils — who have obtained
school bus schedules that typically reduced the number of required
vehicles by 10-20% and miles driven by 12-15% and have an information
management tool for better decision making — and passengers who have
experienced improved service quality.
Underpinning research
UK local councils spend £1 billion every year on school-bus
transportation. For special needs children, where wheelchairs and escorts
are needed and buses can be flexibly configured to hold different numbers
of passengers and wheelchairs, the routing and scheduling of buses is very
complex. A workable schedule must correctly accommodate the child and
their needs, and route the bus to arrive at the address side of the road
and not require the bus to make U turns. A good schedule would seek to
minimise the number of vehicles, and reduce passenger ride-times and CO2
emissions. Research carried out by Julia Bennell, Professor of Management
Science, Management School (1997-) and Chris Potts, Professor of
Operational Research, Mathematics (1986-) has led to the development of
software that allows users to schedule passenger transportation for a
range of complex scenarios. Research was based on a set of methodologies
designed to find good solutions to difficult optimisation problems known
as "local search algorithms". Research extended the state of the art in
routing by designing new ways of representing the road network, and in
optimisation by accommodating a whole range of practical complex
constraints while producing fast solution times and guaranteeing a usable
final solution.
The underpinning research that led to the impact arises from the fifteen
years' experience of developing effective local search methods for a
variety of problem types, and the research undertaken with Logical
Transport (LT), a niche software company, through
- a short LT funded research project (2004) contributing to the
optimisation methodology for the vehicle scheduling component of their
main product, LT Enterprise,
-
Knowledge Transfer Partnership (KTP) in conjunction with Dr Nik
Pearson (Mathematics 2007-2009) (grant 1, 2007-2010) resulting in new
routing algorithms and optimisation methods for the many-to-one
extension of LT Enterprise,
-
Knowledge Transfer Secondment (KTS) (grant 2, 2010-2011) that
generalised and improved the optimisation methods resulting in
Routeforce add-on to LT Enterprise.
Research by Bennell on designing local search algorithms was fundamental
in the work with LT to improve their software. The particular problem
shares common features of sequencing that are used in cutting and packing
[3.1], machine scheduling [3.4] and shipping [3.5]. Her work on problems
with multiple objectives [3.2] directly fed into the more recent work with
LT that provides the facility to select one or more KPIs for optimisation.
Results of the research are presented in [3.6] and [3.3].
Working directly with LT, Bennell and Potts used their research
experience with these methodologies to improve LT scheduling and routing
algorithms in LT Enterprise. During the KTP, Bennell, Potts and Pearson
developed a detailed algorithm for a `many-to-one' vehicle scheduling
problem i.e. passengers are transported from many starting points to one
destination. Significant methodological innovation was required to deal
with problem constraints including a fleet of mixed vehicles types,
variable travel times, flexible vehicle seating for wheelchairs, and
requirements for escorts. The main outcome was a scheduling software
extension to LT Enterprise, (many-to-one) that allows the user to create
and analyse school bus schedules and generate management information for
decision making. For the KTS, Bennell and Potts working with Pearson
(employed since 2009 by LT as a senior researcher) extended the KTP
research to look at day-care scheduling, where there are many pickup
locations and a few destinations. This resulted in the development of the
next generation of scheduling software called RouteForce.
References to the research
Publications
3.1 Bennell J.A. and Song X., A beam search implementation for the
irregular shape packing problem. Journal of Heuristics, 2010, 16, (2),
167-188.
3.2 Bennell J.A., Lee L.S. and Potts C.N., A Genetic Algorithm for
Two-Dimensional Bin Packing with Due Dates, International Journal of
Production Economics, 2013, 124 (2) 547-560.
3.3 Bennell J.A., N. Pearson and C.N. Potts, Drivable routes: Shortest
paths in practice, Discussion Papers CORMSIS-09-15. Under review for
Networks.
3.4 Bennell, J.A., Potts, C.N. and Whitehead, J.D., Local search
algorithms for the min-max loop layout problem. Journal of the Operational
Research Society, 2002, 53, 1109-1117.
3.5 Malliappi, F., J. A. Bennell, C. N. Potts: A Variable Neighborhood
Search Heuristic for Tramp Ship Scheduling. ICCL 2011: 273-285
3.6 Pearson, N., J.A.Bennell and C.N.Potts. Good triangulations yield
good tours, CO 2008 — International Symposium on Combinatorial
Optimization
Grants
Prof Chris Potts and Prof Julia Bennell, KTP: Resource scheduling
algorithms, DTI, January 2007 - Dec 2009, £100k.
Prof Julia Bennell and Prof Chris Potts, KTS: Day care scheduling
algorithms, EPSRC, May 2010 - July 2011, £20k.
Details of the impact
Research by Bennell and Potts and the development of many-to-one and
RouteForce have directly led to commercial gains for Logical Transport
(LT), financial and carbon emission savings for local councils and other
transport providers, and improved experience for passengers.
Benefits to Logical Transport and private industry
As a result of LTs collaboration with Southampton, LT has grown from five
employees and a single product to fifteen employees, with a 58% increase
in turnover between 2009/10 and 2010/11 and moving from a net loss to net
profit [5.1]. Since 2010, LT has used the Many-to-One scheduling software
as part of its core scheduling product [5.2]. This has largely been used
in its consultancy services where LT perform a full analysis of the school
bus requirements for their clients. Four local authorities and two private
schools used this service and the introduction of RouteForce in 2012
attracted two further local authority clients.
Many-to-One and RouteForce have provided LT with the tools to realise a
significant competitive advantage as they can deal with differing needs of
passengers, flexible capacity vehicles and restrictions in the real road
network. Competitors' routing software sees passengers as homogeneous and
fails to take into account practical issues such as arriving on the
address side of the road or the difficulty in making U-turns. The
Southampton-developed tools allow for efficient scheduling of passengers
to vehicles according to any of five key performance indicators: number of
vehicles, mileage, CO2 emissions, vehicle time on road,
passenger time in vehicle. LT can prioritise different KPIs for clients.
[5.3]
LT software that includes the capabilities developed via the
collaborations, are used by a range of clients for transport scheduling,
such as the National Express Hoppa Service (2009) [5.4] for transferring
passengers between hotels and airport terminals. As a result, National
Express are able to provide a fixed schedule service during peak times and
flexible scheduling during off-peak hours. This has delivered improved
customer service in off-peak hours, while reducing operational overheads
and environmental impact. Introduction of the demand response service has
resulted in a 30% drop in expenditure for National Express without
sacrificing customer service. Alan Willson (former Director of Business
Development at LT) said: "(the routing engine developed in conjunction
with the University of Southampton) undoubtedly helped to secure the
deal with National Express."
The financial benefit to LT is quantified in a statement in 2009 by Dale
Howarth, Business Development Manager at LT, said: "Over the past two
years our revenues have increased by over 300%. Our relationship with
the University of Southampton has, and continues to be a significant
contributor to this success". [5.5]
Advantages to local authorities and schools
School transport represents a major expenditure for local authorities and
schools. The pressure on public finances in recent years has meant that
the cost of school bus services has come under particular scrutiny. Since
2010, LT have worked with Bristol, Camden, Devon, Enfield, Halton and
Telford local authorities on school/day care transportation scheduling and
[name of college removed for publication], Trowbridge and St Dunstan's
School on school bus scheduling. London Borough of Camden run a fleet of
11 fully accessible buses for 173 special needs school children, 23 buses
for adult transportation to day care, and 5 demand responsive buses for
adults who cannot access public transport. Each bus costs £100k per year.
The study showed a 20% reduction in vehicles and a 61% reduction in miles
driven (2013). The Transport manager, Callum Johnson, cited the impact of
LT software as one of management information for decision-making including
capacity analysis, and accurate reporting of passengers that is important
for supporting their safeguarding policy. They are about to implement
Routeforce, which he states "will make a massive difference" in their
ability to reduce costs and environmental impact [5.6]. At [name of
college removed for publication], Trowbridge, the study showed a 19%
reduction in vehicles and a 19% reduction in miles driven, with an
estimated minimum cost saving of £100K per annum (2012).
Benefits to passengers
The studies undertaken by LT for the mentioned councils using Many-To-One
and Routeforce have shown higher quality service in terms of ride time
and/or reduced fuel emissions through better routing. School children and
day-care patients benefit from shorter ride times and users of National
Express Hoppa Services benefit from a more responsive service. Since the
software considers seat configuration, passengers that have very special
needs can be accommodated, for example, requiring an escort and vacant
seats next to or in front. This provides a better service for these
children and aids their social integration.
Sources to corroborate the impact
5.1 Statement provided and contact for further corroboration:: Director,
Logical Transport 5.2 Many2One website: http://www.many2one.co.uk/about-logical-transport.html
and Kids 2 School
http://www.many2one.co.uk/online_school_transport_planning_scheduling.html
5.3 Contact: Senior Analytical Consultant, Logical Transport
5.4 Case study, Logical Transport:http://www.logicaltransport.com/hoppa.html
5.5 Presentation by Dale Howarth "How Logical Transport Ltd have
successfully used a Knowledge Transfer Partnership to improve
competitiveness/productivity" at a seminar organised by Solent Synergy Ltd
on "How Local University Expertise can Improve Business Performance", held
at Quob Park, Titchfield Lane, Wickham, 28 January 2009.
5.6 Contact: Transport Manager, Camden Council, London Borough of Camden.