Tackling the global supply chain crisis

Chemotherapy is the most common way to alleviate or mitigate against the spread of cancer. Patients need to go through an exhaustive process with many side effects – SWALOS researchers are working to minimise treatment delays.

“I doubt many people would associate the word ‘cancer’ with ‘mathematical algorithms’ or consider that they can help patients through their chemotherapy treatment most effectively,” says SWALOS team member Dr Melih Celik.

“Chemotherapy day units operate with extremely limited resources and under tight schedules. Designing patient schedules is crucial in determining how efficiently these resources are used – we looked at the scheduling decisions at an outpatient chemotherapy unit of a university hospital in Turkey, where these decisions were previously made using simple rules-of-thumb,” Dr Celik says.

The research takes into account both provider and patient satisfaction. From the provider’s perspective, nurse workloads should be balanced, overtime should be avoided to the best extent possible, and chairs should be highly utilised. Patient satisfaction mainly depends on reducing waiting times due to prior patients’ infusions taking longer than expected.

“These needs require some intricate trade-offs, which conventionally demand building complex mathematical models to find the best set of scheduling decisions. But such models take hours of computing time, which prevents their use at medical facilities where timely decisions are of the essence,” he said.

Time study

The research team conducted a time study of the actual infusion times for more than 200 cancer patients to see how they compared with the times that medical staff estimated would be needed when they set their appointment schedules.

"The actual results showed significant deviations from the estimates. For example, of the 39 patients with an estimated infusion time of 150 minutes, actual times ranged between 61 and 196 minutes. We also observed that shorter infusion durations were underestimated, whereas longer ones were overestimated," he said.

Celik said he and his fellow researchers - Dr. Serhat Gul of TED University, and Nur Banu Demir from Wayne State University, were conscious of the need to create a scheduling model that could be used in situations where time was of the essence, ruling out the use of complex mathematical models that could identify the ’best’ set of decisions but which would take too long to be of practical use in a hard-pressed medical environment.

“We built a simpler approach which not only provides ‘near-best’ schedules in only a few minutes and also outperforms the existing schedules of the chemotherapy unit by a long margin; patient waiting times were improved by 80%, nurse overtime was reduced by more than 30%, and chair utilisation was increased by 5%.

“It is a scene familiar to so many of us – your flight is delayed. Sometimes you might get an explanation – the inbound flight, for example, has been held up, but rarely do we get an understanding of what is at the root of that problem and, more importantly, how can it be fixed? That’s what we’ve done with our study in delay propagation in transport networks,” Dr Giannikas says.

The team, in collaboration with Boeing Company and the Institute for Manufacturing at the University of Cambridge, studied an international hub-and-spoke airline, operating short and long-haul flights from several airports including both domestic and international flights.

The study examined the way in which delays propagate and affect each other, and also the reasons why delay propagation starts and then continues to affect other parts of the air transport network. The researchers drew on data commonly available in airline systems and sought to draw results that can be applied to other hub-and-spoke airlines.

A research method to develop a more holistic view of air transport delays

The case study made use of data extracted by various software systems used by the case airline to manage its operations, such as fleet scheduling, crew management, and passenger itineraries. The dataset covered six months worth of airline operations and includes flight, crew and passenger information. In total the team collected information from over 96,000 flights. The data-driven method consisted of two stages: the first one modelled the airline schedule network and the second extracted delay propagation patterns and root-causes.

"Our data-driven method used a multi-layer airline schedule network perspective which means incorporating aircraft, crew and passenger information in our analysis. Using this method, we were able to develop a more holistic view to air transport delays in contrast to previous studies using only aircraft delay data. Our results show that if we only look at planes and leave out passengers and crew, it is almost impossible to examine – and then fix – delay propagation properly," Dr Giannikas says.

The research showed passenger connections appeared to play a key role in delay propagation, especially for hub-and-spoke airlines that help their passengers move between flights in certain airports.

"So improving airport and airline processes related to passenger connections (passenger and baggage tracking, for example) should reduce the impact of delay propagation," Dr Giannikas says.

The study also identified the importance of effective maintenance and repair of aircraft, as well as the need to improve the ability of airlines and airports to respond to more unpredictable issues, such as severe weather, terrorist threats or industrial strike action.

"An improvement in disruption prediction and recovery capabilities can prepare an airline to deal more efficiently with extreme events, and this might include avoiding network bottlenecks, such as airports and airspace, that can further propagate a delay."

The Covid-19 crisis has put healthcare systems under enormous pressure. Although the number of people hospitalised with Covid is reducing in countries across Western Europe, the current crisis will not be fully stopped until a vaccine is rolled out across the world.

Professor Güneş Erdoğan, a member of the SWALOS team, is developing location, routing, and prioritisation and allocation models to distribute and administer vaccines in an efficient, timely and fair way.

The project will run until the end of March 2023 and will develop models to locate and route medical facilities to administer vaccines, as well as develop scenarios for prioritisation and dynamic simulation to allocate vaccine quantities and vaccination capacity

Research method aims to minimise total mortality

The project combines epidemiological models with resource allocation models in locations covered by NHS Trusts in the UK. This translates to models that estimate the number of infected and deceased based on vaccine allocation decisions. The objective of the model is to minimise total mortality, with a secondary objective of maximising equity – achieving a roughly equal percentage of people vaccinated within each location. The team have collected data from the NHS about hospital capacities, and the infection spread parameters for Covid to feed into the optimisation model. The output is a week-by-week allocation of vaccines to NHS Trusts.

SWALOS has joined forces with Industrial Internet of Things company Logidot to develop technology for improving warehouse operations, cope with demand for ever faster online deliveries and alleviate the stresses placed on the global supply chain.

SWALOS is bringing its algorithm and modelling expertise to the project, while Logidot will further develop technology to track human pickers, forklift trucks, handheld scanners and other hardware at work in warehouses in real time.

“The pandemic has brought our reliance on smooth supply chains to the fore and triggered new trends in warehousing which are all about improving distribution to customers and responding to changing demands,” says Dr Giannikas.

The technology will bring cost savings in traditional warehouse set-ups, as well as aiding the rising trend of urban micro warehouses, and dark stores cropping up in shopping centres and towns across the UK. These new spaces allow companies to fulfil fast deliveries more easily, locating themselves close to customers, with good transport infrastructure, but they are currently often ill equipped to handle high volumes of orders and supply chain fluctuations.

The solution will be based on Logidot's indoor/outdoor tracking technology to continually monitor the movements of material handling vehicles (forklift trucks), equipment (barcode scanners) and order pickers wherever they are in a warehouse.

Very low power wireless hardware have been developed for ease of installation and reduced maintenance cost, alongside innovative dynamic routing algorithms to add constraints (e.g., dispatch times, synchronisation with delivery services and customers pick-ups, zoning, routing and picking optimisation) based on warehouse layout and operational data inputs specific to urban dark stores and micro-fulfilment centres.