To view the full presentation please visit: http://www.theorsociety.com/Pages/Probono/latestnews.aspx
Below is a full write up from the event (by Nigel Cummings)
OR and Oxfam
(The OR Society:
Joint Third Sector SIG, SORG and CORMSIS Event)
A recent presentation given by Tom Cherrett, Associate
Professor, Logistics and Transport Planning, Southampton University, illustrated
the application of remote monitoring technology and optimisation techniques to
collect donations more efficiently.
Many shops around the country have “food banks” where
shoppers can deposit unwanted non-perishable items possibly bought in a BOGOF
deal. Charity shops also often receive items that they cannot sell but are
perhaps too good to throw away. Oxfam recognised that collecting these items
accounts for something like 20% of their income and thought O.R. might be able
to make significant reductions in this.
While the main focus of Tom’s presentation concerned work
done for Oxfam, he also had the time to detail some of the processes involved
in achieving logistics efficiencies in other organisations. Primarily his talk concerned the
implementation of remote sensor technology in donation banks to help
organisations like Oxfam more efficiently judge “fill levels” and allocate
appropriate times for donation bank collections.
Using the fill levels reported daily and derived collection
strategies using tabu search methods Tom Cherrett’s team had achieved
considerable efficiencies and overcome many obstacles associated with donation
bank management. One of the problems they had to address was neatly illustrated
by a slide which showed a small child being used by donation thieves to enter
donation banks and remove contents (Fagin would have been proud!).
Such despicable acts were all too common, and it was a goal
of this research to find ways of reducing this type of pilfering. The work undertaken had involved equipping
Oxfam banks with infra-red sensors that measured how full the banks were,
reporting the data twice daily, and using it to schedule collections more
efficiently whilst reducing thieving.
O.R. methods were employed to help provide indicators as to the
most appropriate times to make collections. The proposed routes output by the
algorithm used for each day of operation were then adjusted by Oxfam's
transport manager to take into account issues such as round balancing, vehicle
access restrictions, staff availability and other constraints.
A trial was run 2013 with the results showing a relatively
modest time and distance savings (~3%) initially, these being limited by
Oxfam's fixed shop servicing constraints. It was estimated that savings of up
to 25% could be made by easing these restrictions. These savings were based on
a set of rules that only allowed banks to become eligible for servicing once
they had reached a specific fill level.
The fill/collection problem was made more complex by the
need for the collection vans to visit Oxfam shops on a fixed schedule basis to
remove unsold textiles too. Following live and simulated trials, the results
suggested that time and distance savings of up to 30% over the current fixed
schedules could be achieved when a minimum bank and shop fill level of between
50% and 60% was used as a collection trigger.
Some of the outcomes from the project are now being developed
commercially in the form of a mobile phone app, to allow area managers, shop
managers and drivers to communicate and manage collection scheduling in a more
dynamic way. Other benefits included greater flexibility for transport managers
working at Oxfam, to schedule ad hoc work
and offer shops additional collections. The Oxfam drivers benefited too as the
efficiencies achieved in logistics and better division of labour meant that
sometimes it was possible to finish work an hour early.
The work with Oxfam is ongoing, and it is expected that
further efficiencies will be achieved.
