Presented by Parvathy Krishnan, Lead Data Scientist

The Geospatial Planning & Budgeting Platform is being developed by researchers and practitioners from Analytics for a Better World Institute in close collaboration with World Bank.

These digital decision-support interfaces focus on promoting inclusive and resilient access to social and economic infrastructure services, including for health. The Geospatial Planning and Budgeting Platform (GPBP) provides users with online, interactive, and collaboration-friendly interfaces for powerful descriptive and prescriptive decision support. The GPBP user interfaces for any given problem statements (or “use-cases”) are presented in two forms:

(i) an interactive website for non-technical end-users,

(ii) a Jupyter Notebook environment that allows for exploratory and background analytics and visualization to be made more transparent and accessible.

In this talk, the overview of two GPBP use cases – Health Facility Location Optimisation in Timor Leste and Stroke Case Accessibility Optimization in Vietnam – will be explained. The challenges and opportunities of this platform along with potential extensions and the plans for the way forward will also be discussed.

1 JULY 2022 WEBINAR RECORDING

Presented by Prof. dr. Ger Koole (Vrije Universiteit Amsterdam)

In this talk, I will make a comparison between call centers and hospitals, from the point of view of operations management.

I will show how advanced prediction and optimization techniques impact call center operations and how academia and industry interact to move the field forward. We will compare this with the current state of our health care system, identify similarities and differences, and discuss what blocks further improvements in quality and efficiency.

About the speaker:

Ger Koole is full professor of business analytics and optimization at Vrije Universiteit Amsterdam. He is also founder and lead scientist of the call center workforce management company CCmath. He obtained his PhD at Leiden University in 1992 with theoretical work on the control of queueing systems.

Before moving to Amsterdam he held the postdoc positions at CWI and INRIA. Over the years, his interests moved to applications of optimization under uncertainty, especially in call centers, health care and revenue management.

Presented by Dr. Jason Atkin, Associate Professor (School of Computer Science, University of Nottingham)

Dr Jason Atkin is a member of the Computational Optimisation and Learning lab in the School of Computer Science at the University of Nottingham. He has been working with NATS and Heathrow airport since he started his PhD on the topic in late 2003. Although he has also worked on various other airport optimisation and modelling problems, his work has primarily been focusing on real-world runway sequencing at the airport, considering the ‘often messy’ real world constraints and characteristics of the problems. Having previously been a software engineer in industry, he applied various software engineering techniques to the challenge, which gives a somewhat different perspective to the usual academic approaches.

This talk will discuss the work with NATS and Heathrow, which contributed to Heathrow being able to predict take-off times accurately enough to gaining CDM compliance. Various versions of the problem will be discussed, introducing both the real world situation at Heathrow and considering the various academic simplifications that have sometimes been used for similar problems, and their implications in the real world. Ongoing work with Heathrow is still providing further enhancements and improvements over time.

Presented by Dr. Anne Liret, Research Manager (British Telecommunications)

In the face of the climate emergency, several measures are being taken to move to more sustainable sources of energy. The transport sector has increasingly adopted Electric Vehicles (EVs) in order to cut down on greenhouse gas emissions. This has led to the necessity of accounting for these technologies in the area of field services delivery. In particular BT Research have been developing technical solutions to support the introduction of EV into their large fleet operations. This led to new models to plan for new EV chargers deployment minimising the energy risk, and then to schedule jobs and charge activities of electric vehicles in the feasible operational manner for field engineers while minimising the productivity impact.

By optimising the infrastructure, the required investment can be determined as well as minimising the disruption to the operations caused by the rollout of the new VE fleet and infrastructure can be minimised. As usual in operational research applied to decision-making support, the measuring of the impact on real-case trials is key to make sure changes adoption is going smoothly.

1 APRIL 2022 WEBINAR RECORDING

Presented by Alex Fleischer, Optimization Expert (IBM)

When trying to solve puzzles, practitioners train themselves on OR techniques (and other techniques). Puzzles are good ways for large and small companies to have OR practitioners from academia and consulting take a look at their specific issues (ROADEF / Euro challenges). Puzzles can support challenging students to show motivation and skills.

Why puzzles are very interesting with regards to equivalent computational challenges? We need to map real world concepts to mathematical concepts and that’s useful! They also tell a story, so they’re very easy to share and explain.

Kaggle challenges are part of the buzz around data science. During the presentation, I’ll mention other public challenges with examples: The IBM Ponder this challenge, the Decision Management challenge, and, not to forget, the “Comité International des Jeux Mathématiques”, and I will request the audience’s feedback with regard to deciding if challenges are good training paths for OR consulting and, also, if real-world business OR helps practitioners get better at challenge

4 MARCH 2022 WEBINAR RECORDING

Presented by Ugur Arikan (Operations Research Scientist, Deutsche Post DHL) and Jonas Witt (Operations Research Scientist, Deutsche Post DHL)

Traditional last-mile delivery planning purely based on optimization methods often lacks important real-life aspects and thus does not satisfy relevant operational requirements. Only experienced couriers have tacit knowledge about the delivery area and its customers, forcing them to deviate from planned routes. They know where to find good parking spots, when to best approach certain business customers, and which neighbors to approach at what times when the actual recipient is not at home. This tacit knowledge is almost impossible to collect and maintain, let alone to incorporate in optimization algorithms.

Thus, we at Deutsche Post DHL developed an algorithm following a different approach: We aim at implicitly learning about this tacit knowledge from historical tours and combine this with optimization algorithms to plan routes that an experienced courier would choose. In this talk we will present details of our algorithm, which incorporates machine learning, statistics, and optimization in a novel way. Furthermore, we show how it impacted last-mile delivery planning at Deutsche Post after its rollout across Germany.

Presented by Moulay Driss El Alaoui Faris, Energy Decision Scientist (Air Liquide)

Embedding OR functionalities in operational tools to address a large class of problems requires designing the software architecture to provide the required modularity.

In this talk we will present comparative design of component primitives (Nodes Arcs Boxes) in Python and Julia. A particular exploration question is how to use Julia for the design of object oriented software.

14 JANUARY 2022 WEBINAR RECORDING

Presented by Çağrı Doğuş Iyican (Trendyol Express) and Yahya Ertuğrul Geçkil (Trendyol Express)

Trendyol Express (TEX) is the sub-branch in Trendyol that leads daily cargo operations. As in other cargo cases, it needs depots for storing, sorting and transporting to meet customer demand and manage seller supplies. In this context, TEX has two kinds of depots for its operations which are sorting centers and cross-docks (x-docks). Sorting centers are responsible for taking packages from sellers and delivering these packages to x-docks whereas x-docks are responsible for delivering packages to customers. In a growing market, first mile and last mile demands in e-commerce forces planners to design an efficient network in order to optimize delivery times. This project aims to find the optimal locations for sorting centers and x-docks under several operational constraints.

Increasing e-commerce demand is hard to satisfy because of delivery bottlenecks. Cargo capacity growth ratio is far below e-commerce sales growth. Also, delivery costs have a major effect on expenses for e-commerce companies. Furthermore, delivery management is crucial for customer satisfaction levels. However, as well as customers, the number of sellers taking place in marketplaces is also increasing dramatically. Since it is not possible to find out best practice for tens of thousands of sellers manually, it is quite possible that resources are wasted, customers can not reach their orders in desired time and sellers deal with unnecessary costs. Besides, managers waste weeks only to find out a feasible plan. A mathematical model is developed for solving the dispatching problem for marketplace sales beside backlog amount. The model aims to distribute backlog and sales by minimizing the total delivery time.

Professor Gilberto Montibeller (Loughborough University, UK; University Southern California, USA),

Professor L. Alberto Franco (Loughborough University, UK; University del Pacifico, Peru)

Emerging health threats, such as the COVID-19 pandemic, create extensive health, economic and social problems. A key challenge for health experts and policy makers is deciding how to balance and reduce the risk of these threats.

In this applied research project, we developed an advanced framework to support the prioritisation of emerging health threats. The project won the prestigious EURO Excellence in Practice Award conferred by the Association of European Operational Research Societies, during the 31st EURO Conference on Operational Research held in Athens this year.

The research project achieved the following impacts in two global organisations:

(i) enhanced the quality of health experts’ policy recommendations at the UK Department for Environment, Food and Rural Affairs, and improved health risk control regulations;

(ii) informed the development of new international standards for the Food Standards Joint Programme of the Food and Agriculture Organization of the United Nations and the World Health Organization.

The contributions to Operational Research arising from the ongoing research programme are built upon three themes, which are focused on improving decision capability for the prioritisation of emerging health threats. These are tool development, process enhancement and competence building:

• Tool development encompassed the creation of rigorous decision analytic models to support decision making in the prioritisation of emerging health threats, as well as the design of risk management support tools that enable policy makers to use these models.

• Process enhancement focused on the redesign of decision processes to address the challenges posed by the prioritisation of emergent health threats, to enable the embedding of risk management support tools within organisational routines.

• Finally, competence building supported the development of health experts’ decision and risk analysis skills, together with the effective deployment of value-focused decision making and sound health risk management practices.

Health security is an important area of application for Operational Research and in this talk we will share our experiences in conducting these challenge and successful applied research projects.

5 NOVEMBER 2021 WEBINAR

Presented by Professor Goos Kant (Professor of Logistic Optimization, Tilburg University; Managing Partner, ORTEC)

Thanks to all great developments in mathematics and computing power, we are able to solve more complex problems within a reasonable amount of time. Machine learning and other AI-techniques can boost the power of OR. But are we also able to explain the outcome to the industry? Optimal is not always logical, which leads to challenges both in modeling and in implementation. In this presentation I like to share my thoughts, based on my professorship “Logistic Optimization” at Tilburg University and my long experience (in parallel) as “Optimization Evangelist” at ORTEC. ORTEC is a global and leading partner in data-driven decision support.

8 OCTOBER 2021 WEBINAR