2020 Edelman Competition
Join us in celebrating the 49th annual Franz Edelman Award for Achievement in Advanced Analytics, Operations Research and Management Science, the world’s most prestigious award for achievement in the practice of analytics and operations research.
This year's five finalists included: Carnival, Deutsche Bahn, IBM, Intel, and Walmart.
Congratulations to Intel for their 2020 award winning submission and on being named by the prestigious panel of judges as the recipient of the 2020 INFORMS Franz Edelman Award!
Watch the Finalists Presentations
Joint revenue and inventory optimization for multiple heterogeneous brands in the world’s largest cruise company
In 2020, more than 32 million people will cruise on over 278 ships, with 14 million guests from North America, 6.7 million from Europe, 4.2 million from Asia, and 1.5 million from Australasia. Of those guests, nearly 50% sail on Carnival Corporation & plc brands. As the world’s largest leisure travel company, Carnival Corporation & plc provides guests around the world with extraordinary holiday experiences and exceptional value.
Carnival Corporation & plc is comprised of Carnival Cruise Line, Princess Cruises, Holland America Line, Seabourn, P&O Cruises (Australia), Costa Cruises, AIDA, P&O Cruises (U.K.), and Cunard Line, with operations in North America, Australia, Europe, and Asia, and together serve tens of millions of guests on cruises every year.
Carnival Corporation & plc undertook a cross-brand revenue management diagnostic that revealed opportunities to strengthen the use of operations research, innovation, and automation. A cutting-edge revenue management system was needed, however, existing solutions from the airline and hospitality industries were not compatible with the idiosyncrasies of the cruise domain. As such, the company partnered with Revenue Analytics to build a complete revenue and inventory system from scratch.
Business leaders, operations researchers, and software engineers came together from around the world to build YODA (Yield Optimization and Demand Analytics). Over four years, they built a system founded on advanced analytics to forecast demand, measure price sensitivity, and ultimately recommend pricing and inventory controls.
Prior to YODA, each of the Carnival brands was managing revenue and making decisions using its own tools and processes. Tools that existed across the brands varied significantly from simplified forecasting and cancellation predictions, to analysts making decisions in Excel tools by comparing to a similar cruise from history. These tools were capable of detecting large, sudden deviances from expected behavior, but were less adept at detecting slowly-developing patterns or handling quicker incremental adjustments needed to maximize revenue.
One of the initial barriers to overcome was that of geographical diversity; organizational structures and time zones are real barriers to working across brands within the corporation. The YODA project was one cross-brand team, with defined roles and responsibilities for members based across Europe, Australia, and the U.S. In order to leverage the scale by combining efforts across brands, the team first needed alignment. Overall, the team identified 80 different categories and questions to resolve and align on an approach before designing and building a single system that could be consistently used by all brands, which were consolidated into a “Brand Alignment Matrix.” The process took almost three months, but once complete, set up the remainder of the project for success.
YODA is a system that leverages a unique quadratic programming optimization formulation (jointly optimizing both price and inventory allocations across multiple time periods). The inputs come from several machine learning algorithms to predict demand combined with an elasticity model derived from an exponential curve fit to represent the unique price-sensitivity behavior observed in the cruise industry. These models generate millions of price recommendations a day to price 65 ships.
Not only does Carnival generate 1.5% to 2.5% uplift in net ticket revenue as a result of YODA, according to pilot testing, but it also elevates the role of the pricing analyst in making strategic decisions, facilitating cross-brand coordination, and quickly reacting to a dynamic industry. As an additional measure of success, since initial deployment YODA has been adopted by an additional brand indicating it continues to be viewed very favorably internally.
In addition to the revenue benefits, YODA has driven renewed strategic conversations around the revenue management teams in each of the brands, by giving the science teams and analysts the ability to challenge long-standing principles of cruise revenue management using YODA outputs and associated models.
YODA has also set the new standard for cross-brand projects in Carnival Corporation. Not only did the project complete on time and under budget, it also fostered a level of collaboration never before seen between the Carnival brands.
Carnival Corporation & plc
Carnival Corporation & plc is the world’s largest leisure travel company, providing guests around the world with extraordinary holiday experiences and exceptional value. With operations in North America, Australia, Europe, and Asia, its portfolio features Carnival Cruise Line, Princess, Holland America Line, Seabourn, P&O Cruises (Australia), Costa, AIDA, P&O Cruises (U.K.), and Cunard.
Together, the corporation’s cruise lines operate 105 ships with 254,007 lower berths visiting more than 700 ports around the world with 15 new ships scheduled to be delivered through 2025. Carnival Corporation & plc also operates Holland America Princess Alaska Tours, the leading tour company in Alaska and the Canadian Yukon.
Traded on both the New York and London Stock Exchanges, Carnival Corporation & plc is the only group in the world to be included in both the S&P 500 and the FTSE 100 indexes. Carnival Corporation employs a talented, passionate, and diverse workforce of over 150,000 people from nearly 150 countries, and its brands host more than 13 million guests annually – about half of the overall global cruise market. Combining over 265,000 daily cruise guests and 100,000 shipboard employees, more than 365,000 people are sailing aboard the Carnival Corporation fleet every single day totaling about 93 million guest cruise days per year.
Revenue Analytics is a SaaS company that helps big companies make big revenue decisions in pricing, products, and promotions. Our analytics solutions drive millions in revenue uplift and eliminate wasted time.
Founded in 2005, we leverage our deep experience in pricing and revenue management to empower our customers with predictive analytics engines that drive complex pricing decisions.
These capabilities have been deployed across many industries with Revenue Analytics being responsible for adjusting the prices on more than 40% of North American hotel rooms nightly, pricing 45% of all passenger cruise trips globally, and have over 42% of U.S. radio ad sales running through our platform.
On the Right Track: Deutsche Bahn Schedules Train Rotations Using Hypergraph Optimization
Deutsche Bahn (DB) uses novel operations research methods to optimize the rotations of its rolling stock in its cargo, regional, and long-distance passenger transport divisions. Advanced optimization methods allow DB to manage timetable changes in order to offer more and better services to its customers, cut greenhouse emissions, and enhance regularity. DB transports about 5.6 million people each day on trains throughout Germany, as well as 700,000 tons of goods. These are carried by over 1,400 long-distance trains, 22,000 regional passenger trains, and 2,800 cargo trains, which are operated by a fleet of 3,800 locomotives, 4,500 passenger coaches, more than 80,000 freight wagons, and over 4,000 railcars, among them 274 ICE high-speed trains.
With a constantly increasing number of trains and a continuously fluctuating demand on a limited infrastructure in a deregulated, competitive railway market, it becomes harder and harder to schedule these services. It is difficult to direct the trains through bottlenecks and around construction sites to avoid cancellations, deadhead trips, and orientation mismatches. To address the challenges of using its rolling stock at full capacity, Deutsche Bahn drafted a vision of reinventing railways: “Strong rail” is DB’s overarching strategy to become more efficient in operations, develop a more powerful organization, and stimulate innovations. The cornerstone of this plan is the development and implementation of a new generation of advanced operations research methods, to first and foremost schedule DB’s core asset: its rolling stock of locomotives, wagons, and railcars.
To this purpose, DB launched the development of a new train/locomotive rotation optimization system called FEO/LEO in 2008. In order to provide a company-wide planning platform for three divisions, the solution was implemented as a modular system around adaptable optimization kernels. These kernels were flexibly customized to suit a large variety of planning tasks. FEO/LEO is directly linked to DB legacy planning IT, such that a seamless workflow is ensured. A particular strength in FEO/LEO is an integrated feedback algorithm that allows planners to approach an aspired solution by applying minimal changes with respect to the planning constraints. This had a profound impact on the scheduling methodology at DB, which changed from constructive to analytical planning.
The FEO optimizers are based on completely new and sophisticated mathematical methods of algorithmic network and hypergraph theory. This research is done together with Zuse Institute Berlin and LBW Optimization. The novel approach allows for an integrated treatment of the main operational requirements all within one generic model. A novel coarse-to-fine algorithm allows the solution of large-scale instances with up to 100 million variables. The approach also gives rise to deep theoretical developments in the mathematical foundations of operations research.
FEO has been operating since 2013 and has increased revenues in long-distance passenger transport by more than $26 million per year by reducing cancellations, and by $25 million per year in regional passenger transport by successful bids for tender contracts.
In the cargo division, LEO went online in 2018 with estimated savings of over $29 million per year. Moreover, dangerous and arduous labor is reduced with less couplings. Resource planners benefit in their day-to-day work from the powerful scheduling functionality of FEO/LEO, which reduces repetitive tasks and allows room for analyzing different options.
FEO/LEO is a first and key step toward the implementation of DB’s strong rail strategy. Driven by advances in operations research, strong rail will provide fast, comfortable, and efficient transport with minimal environmental footprint.
Deutsche Bahn AG (DB)
Deutsche Bahn AG (DB) is the largest European railway company. It is divided into several companies, including the rail transport service providers DB Fernverkehr (long-distance passenger), DB Regio (regional passenger), and DB Cargo (freight). DB Fernverkehr provides regular national and international daytime services. To ensure this, it operates a close network of almost 900 daily long-distance rail services in Germany.
Apart from national products, each day cross-border connections to neighboring countries are on offer – mainly in cooperation with partner railways. This enables DB Fernverkehr with its ICEs, ICs, and ECs to carry approximately 340,000 customers to their destinations every day. DB Fernverkehr has more than 16,500 employees. DB Regio is market leader in German regional public transport. It serves about seven million people in over 22,500 trains per day. It employs about 800 locomotives, 4,000 train cars and has over 35,000 employees.
DB Cargo is the largest European rail freight company. It operates most freight stations with the largest fleet on the continent, consisting of over 3,000 locomotives and 80,000 cars. Almost 60% of all transports are cross-border traffic. It serves both wagonload freight and single car freight all over Europe. The network goes from Lisbon (Portugal) over Nishnij-Novgorod (Russia) up to Shenyang (China).
Zuse Institute Berlin (ZIB)
The Zuse Institute Berlin (ZIB) is an integrative research institute for application-oriented mathematics and data-intensive high-performance computing. Its research focuses on modeling, simulation, and optimization with scientific cooperation partners from academia and industry in application areas ranging from drug design to flight planning. ZIB’s services to the scientific community include the operation of a tier two supercomputer and the BRAIN research network, research campus MODAL, Berlin-Brandenburg library cooperation, and the digis digitization office of Berlin. Named after German computer pioneer Konrad Zuse, the Institute is located on the campus of Freie Universität Berlin. ZIB is closely linked to the three Berlin universities and part of Berlin University Alliance. The institute has about 220 employees.
LBW Optimization GmbH (LBW)
LBW is a leading provider of consulting and optimization services for public transport operators, railways, airlines, and traffic management companies. It excels in numerical algorithms for the solution of large-scale optimization problems, in particular, for vehicle and crew scheduling applications, and for problems in network design and operation. LBW has strategic partnerships with leading suppliers of planning systems that bring LBW technology to customers all over the world. MCF, LBW’s open source network simplex algorithm, is a long-standing component of SPEC’s CINT CPU benchmark test library.
Predictive Analytics for Server
Incident Reduction in the Wild
While the Internet has become a major backbone of our society, the computers supporting this Internet can and will break.
IBM creates value for clients by providing integrated solutions and products that leverage data, information technology, deep expertise of industries and business processes, with trust and security and a broad ecosystem of partners and alliances.
IBM solutions typically create value by enabling new capabilities for clients that transform their businesses and help them engage with their customers and employees in new ways. These solutions draw from an industry-leading portfolio of consulting and IT implementation services, cloud, digital and cognitive offerings, and enterprise systems and software, which are all bolstered by one of the world’s leading research organizations.
Among IBM’s many businesses, Global Technology Services (GTS) is the main focus of our work. IBM GTS operates and manages some of the world’s largest data centers, and thereby some of the world’s most mission-critical workflows and franchises. GTS helps clients along their journey to the hybrid cloud, leveraging the best of their existing systems in the context of the regulatory, security, and workflow of their industry.
Every day, GTS operates many thousands of computers for IBM clients. To help them manage this, GTS and IBM Research have developed and deployed a system that identifies and proposes deep analytics and operations research (O.R.) remedies to keep IT systems operational. This Predictive Analytics for Server Incident Reduction (PASIR) solution developed at IBM has been broadly deployed to more than 360 IT environments since 2013. These environments, covering every sector from banking to travel to e-commerce, are serviced by IBM support groups.
More specifically, today, incidents occurring on servers, including the descriptions of the problems and the resolutions, are documented into an account-specific ticket management system. PASIR first classifies the incident tickets of an IT environment to identify high-impact incidents describing server outage and performance degradation issues by using the respective ticket descriptions and resolutions. Second, the occurrence of these high-impact tickets is correlated with server properties and utilization measurements to identify troubled server configurations and prescribe improvement actions through statistical multivariate analysis and simulation. In this contribution, we present the findings from deploying our machine learning solution in the wild.
We describe the PASIR methodology, from ticket classification to the recommendation of modernization actions. We also demonstrate the model effectiveness by comparing predictions on the impact of prescriptive actions with actual system improvements. We have applied PASIR to more than 840,000 servers since 2013 resulting in more precise refresh spending and environment stability, saving our clients an estimated $7 billion since that time.
IBM produces and sells computer hardware, middleware, and software, and provides hosting and consulting services in areas ranging from mainframe computers to nanotechnology. IBM is also a major research organization, holding the record for most U.S. patents generated by a business (as of 2020) for 27 consecutive years. Inventions by IBM include the automated teller machine (ATM), floppy disk, hard disk drive, magnetic stripe card, relational database, SQL programming language, UPC barcode, and dynamic random-access memory (DRAM). The IBM mainframe, exemplified by the System/360, was the dominant computing platform during the 1960s and 1970s.
IBM has continually shifted business operations by focusing on higher-value, more profitable markets. This includes spinning off printer manufacturer Lexmark in 1991 and the sale of personal computer (ThinkPad/ThinkCentre) and x86-based server businesses to Lenovo (in 2005 and 2014, respectively), and acquiring companies such as PwC Consulting (2002), SPSS (2009), The Weather Company (2016), and Red Hat (2019). In 2015, IBM announced that it would go “fabless,” continuing to design semiconductors, but offloading manufacturing to GlobalFoundries.
Nicknamed Big Blue, IBM is one of 30 companies included in the Dow Jones Industrial Average and one of the world’s largest employers, with over 350,000 employees (as of 2018) known as “IBMers.” At least 70% of IBMers are based outside the United States, and the country with the largest number of IBMers is India.
IBM Research is the innovation engine of the IBM Corporation. It is the largest industrial research organization in the world with more than 3,000 researchers in 12 labs across six continents. IBM Research plays the long game, investing now in tomorrow’s breakthroughs. IBM’s scientists are charting the future of AI, breakthroughs like quantum computing, how blockchain will reshape the enterprise, and much more. IBM is dedicated to applying AI, analytics, and science to industry challenges, whether it’s discovering a new way for doctors to help patients, teaming with environmentalists to clean up our waterways or enabling retailers to personalize customer service. Scientists from IBM Research have produced six Nobel Laureates, 10 U.S. National Medals of Technology, five U.S. National Medals of Science, six Turing Awards, 19 inductees in the National Academy of Sciences, and 20 inductees into the U.S. National Inventors Hall of Fame.
IBM Global Technology Services
IBM Global Technology Services designs, builds, and runs the foundational systems and services that is the backbone of the world’s economy. IBM Services partners with the world’s leading companies to build smarter business by reimagining and reinventing through technology, with its business insights, industry-leading portfolio and world class research and operations expertise leading to results-driven innovation and enduring execution. IBM’s experts in business, technology, and industry use advanced technology to help clients reduce cost and risk, achieve compliance, accelerate speed to market, create new revenue streams, and establish a security-rich and reliable infrastructure that’s ready for AI and hybrid cloud. IBM Services’ clients represent the cornerstones of their industries and include 4 of the top 5 airlines by revenue, 8 of the 10 leading mobile operators, 8 of the 10 largest automobile manufacturers, 4 of the 10 largest global retailers, 7 of the 10 biggest insurance companies, and the world’s 10 largest banks are among IBM’s assets.
Intel Realizes $25 Billion by Applying Advanced Analytics from Product Architecture Design through Supply Chain Planning
Intel’s products support a wide variety of customers spanning a broad spectrum of usage models each with a specific set of product requirements. The requirements differ with respect to number of compute cores, processor speed, power consumption, rapid memory access, and input/output ports. Efficient product feature design integrated with supply chain planning is critical to Intel’s success given its scale, complexity of its products and manufacturing processes, and the highly capital-intensive nature of the semiconductor business.
At any point in time, Intel is delivering hundreds of different products to its customer base through its large and complex manufacturing and supply chain while at the same time designing the next generation of products to meet its customers’ needs. In response to an exponential increase in complexities, Intel developed an innovative set of capabilities using advanced analytics that span product feature design through supply chain planning with the goal of maximizing revenue while minimizing cost wherever possible. This set of corporate-wide capabilities is fast and effective, enabling analysis of many more business scenarios in much less time than previous solutions while providing superior results including faster response to customers.
A large part of the product feature design problem is defining a finite number of stock keeping units (SKUs) with associated values for key features to best cover market needs and balance against manufacturability and costs. Given the broad range of Intel’s application areas and customers, the combinatorics faced by product feature designers is daunting and involves search spaces of plausible plans numbering from 1010 to 1050.
The output of the product feature design process feeds the supply chain planning and manufacturing execution process, which routes products through 6+ manufacturing stages, at 200+ internal and outsourced factories across 13 countries. Intel’s manufacturing flow begins with wafer fabrication. Each wafer is a thin circular disk of silicon 200-300mm in diameter and containing 15 to 15,000 complete integrated circuits, or die, depending on die size.
Each wafer is separated into a number of die that are placed into a protective container with external connectors, which will enable attachment to printed circuit boards. At times multiple die from different wafers are used for multi-chip products. Intel is especially focused on the efficiency of using every die on every wafer to supply a saleable product across the architectures of the entire Intel product portfolio.
The assembled items are then tested to measure actual performance. Tests help separate items into categories defined by performance ranges for a combination of the key features. Lastly, the items go through a Finish process that results in individual SKUs. Each wafer could have hundreds of thousands of possible manufacturing routes before it ends up as one of the SKUs shipped to customers. Faster, better, and more integrated decision making is increasingly important to Intel’s decision process, as its products have become more complex.
Each year, Intel introduces several new products and plans for roughly 400 unique wafer types totaling 2.5 million of wafer volume, which results in 4,000 unique SKUs accounting for 600 million in SKU volume. Products can flow through six or more manufacturing stages at more than 200 internal and outsourced factories across 13 countries.
Given the scale of Intel, the complexity of our products, and highly capital-intensive nature of semiconductor manufacturing, efficient product feature design integrated with optimal manufacturing processing and routing translates to billions of dollars of benefit per year, realizing increased revenue by an average of $1.9 billion and reduced cost by $1.5 billion annually, for a total benefit of $25.2 billion to date.
It also supports Intel’s sustainability efforts, enabling Intel to contribute to its environmental protection goals. Semiconductor manufacturing has a high reliance on water and Intel is leading the way with respect to reduction in water usage and the resultant waste water. Intel recycles the majority of the water it uses for manufacturing. The post-manufacturing treatment practiced by Intel produces water that often exceeds local government drinking water standards. Intel’s supply chain planning capability resulted in the reduction of water usage by 2 billion gallons of water and more than 500 million gallons of waste water prevention so far over the 10-year timeframe of the project.
Recently, Intel-driven computers have enabled the advent of Cloud Computing, Big Data, and the practical resurgence of Artificial Intelligence. Equally important has been the way Intel has achieved these technical and economic results.
We are a world leader in the use of green power and conflict-free minerals, and we have an ongoing focus on reducing water pollution and greenhouse gas emissions.
A Multi-Objective Price Optimization Framework in Stores Using Reinforcement Learning
Inventory optimization is one of the core merchandising operations for retailers. A key objective of the inventory optimization workstream is to reduce inventory on or by a specific date, often times through price reduction. Typically, this is done to make space for new items, reduce inventory of overstocked or seasonal items, or move perishable items such as meat or milk – but there are other reasons as well. As the world’s largest retailer and grocer, Walmart strives to find the optimal price point for its general merchandise and groceries. However, when inventory reduction becomes necessary it traditionally implemented a price reduction strategy, which may require up to three different pricing changes, to help move unsold inventory from its stores.
Inventory reduction is a challenge for every retailer, but Walmart’s size and scale means the challenges, opportunities and costs are greater than other retailers:
- More than 4,750 U.S. retail stores, including more than 3,500 Supercenters that can carry more than 125,000 SKUs each
- More than 800 product categories across Grocery, Health and General Merchandise, of which more than 200 categories experience price reductions each year
- Shelf space containing slow-moving or unsold merchandise needs to give way to opportunities to stock them with better-selling items, which impacts revenue and increased costs
- Merchant’s input into the suggested price reductions needs to be taken into consideration as it increases the complexity of their inventory optimization operations and impacts anticipated revenue.
With Walmart’s previous strategy, the time and labor cost required to re-label discounted merchandise up to three times – across all stores and product categories -- was substantial. Determining reduction strategies had to be done on a store-by-store basis, as not all stores had the same inventory optimization issues. Afterwards, the effort to remove and replace merchandise was also significant. In order to maximize sales revenue and reduce operating costs, Walmart created an intelligent algorithm that accelerates sales of time-sensitive merchandise across individual stores, categories and SKUs by optimizing price reductions and minimizing the associated costs of relabeling and removing time-sensitive goods.
Walmart’s inventory optimization algorithm ingests data from individual stores including aggregated sales data and operating costs, how much and what types of merchandise to reduce, and the dynamic time frame for when the merchandise must be sold in order to make way for new merchandise. The core algorithms used to determine the repricing policy originate from mathematical optimization and deep reinforcement learning techniques. This approach applies data analytics, reinforcement learning, and dynamic optimization to make automated decisions (e.g., when to initiate price reductions and by what percentage price reduction percentage of markdown, when to give the markdown price) for each individual product at each store. This results in a high-performance model and a price adjustment policy tailored to each store.
Using machine learning to find the single best optimized price for merchandise and grocery items, Walmart is now able to successfully sell up to 80 percent of time-sensitive items with its initial price reduction, rather than going through the processes three separate times. This has resulted in lowered operating costs and increased sales of these designated items. This has resulted in lowered operating costs and increased sales, with some stores experiencing up to 15% higher sales of these designated items.
Freed from the constant task of re-labelling items, Walmart associates are able to perform higher-value tasks, such as helping more customers and providing better service. The algorithm has also had a significant impact on the daily lives of about 3,000 merchants by shifting responsibilities from repetitive, manual data-entry and manual decision-making on price fluctuations to handling exceptions and strategic scenario planning.
Walmart’s new inventory reduction protocol was implemented in early 2019 and has achieved outstanding sell-through rate and cost-saving performance. The system has successfully reduced excess inventory while boosting sales of that inventory, resulting in significant cost savings. This also improved customer purchasing power and reduced associate workload and related costs in stores. The operational savings and efficiency gains from Walmart’s inventory optimization algorithm gets passed along to customers in the form of its Everyday Low Prices, so customers get the right product at the right price every time.
Thank You to our 2020 Coaches
|Irv Lustig, CAP
Coaching Deutsche Bahn
New Jersey Institute of Technology
OTIS Elevator Company
European University Viadrina
|Aaron Burciaga, CAP
|Harrison Schramm, CAP
Thank You to our 2020 Edelman Judges
Pooja Dewan, 2020 Chair
OTIS Elevator Company
University of Arkansas
University of Connecticut
University of Dayton
Bristol Myers Squibb