Some of the most complex problems companies must solve don't involve massive amounts of data, but the use of constraint based analytics.
When people think about cloud computing, they often think about big data – pulling insights out of copious amounts of information. But there’s another side to the potential offered by cloud-based analytics: the opportunity to solve highly complex problems with a large number of competing variables quickly and cost effectively.
The reality is that some of the most complex problems companies must solve don’t involve massive amounts of data at all. Scheduling is a great example of this. Whether an organization is looking to schedule events, plan bus routes, or optimize package delivery – the actual data involved is often relatively discrete to work with. The challenge arises from the complex factors that need to be considered in order for an organization to identify the optimal schedule given their situation and critical priorities. The right schedule can help organizations reduce costs, improve efficiencies, and even expand their reach while not compromising on quality.
While ten years ago, it might have been impossible to move beyond the basics when it comes to scheduling priorities, this isn’t the case any longer. Numerous organizations globally are shifting their mindset around scheduling. Rather than looking to schedule generation as an obstacle, these organizations are finding ways to optimize their scheduling process across a complex array of factors. Using innovative technologies and solutions like custom algorithms and distributed computing, these organizations are beginning to identify and analyze different scheduling options to determine the ones that best align with their priorities. By doing this, these organizations are creating a distinct competitive advantage.
Scheduling optimization involves the use of constraint based analytics to help organizations identify the best possible schedule according to their objectives. Under this model, organizations can assign values to both scheduling objectives and challenges they want to avoid. For example, a delivery company might want to maximize their overall speed of delivery, but need to prioritize the delivery of high priority packages when developing schedules.
While this sounds like a complex undertaking, organizations can use flexible distributed computing based solutions to develop and assess the trade-offs between different scheduling options in order to make the process accessible and cost effective. Through these types of solutions, organizations can become far more efficient at scheduling, while also generating more value.
Major League Baseball (MLB) has arguably the most complex scheduling challenges of all major sports leagues. The MLB’s 2,430 game schedule includes almost twice the number of games played per year as other large sports leagues. This translates to every team playing 162 games in just 186 days. When the MLB engaged KPMG to assist with its scheduling improvement initiative, it was not just the amount of games that created complexity, but the need to consider other variables, including achieving new benchmarks in broadcasting, spacing of rest days, efficiency of travel, and protecting opening day games against rainouts.
KPMG’s scheduling optimization team worked with the MLB to develop the 2018 schedule utilizing our proprietary cloud-based optimization engine, which was customized to meet their unique needs. After working with the MLB to understand the subtleties of their challenge, we had the scheduling application running on as many as 400 computing processors to generate and evaluate trillions of possible schedules. This allowed us to efficiently provide the MLB with the most optimized schedules for review with the 30 MLB teams.
“One of the things that makes the MLB’s schedule unique is that it’s built around a series of repeating games,” explains Chenlu Lou, a Senior Analytics Specialist with KPMG in the US. “To solve this in the most effective way, we built a daily expansion procedure on top of our proprietary optimization engine. This allowed us to first generate a schedule at the series-level, and then expand it to the daily level.”
The 2018 MLB schedule will improve key performance indicators, while also making travel itineraries more efficient. Over time, it is expected that further efficiencies will be achieved as additional priorities and information are integrated into future scheduling processes.
In order to be successful, organizations need to know that they are operating efficiently, cost-effectively, and in a way that improves their customer outcomes. While it might be a ‘small data’ problem, sorting through the high levels of complexity can be critical for organizations to achieve their strategic objectives.
No organization is immune from the potential benefits. Professional and amateur sports leagues working to create crowd-pleasing and fair schedules, universities and colleges working to schedule courses, transit authorities planning bus routes, hospitals arranging nurse scheduling, and even telecoms working to schedule installation or repair visits can gain substantial value. Using scheduling optimization, any organization can become more efficient at scheduling, while ensuring that a broader range of objectives and challenges are integrated and evaluated as part of the scheduling process.