The faster planning system developed by MIT researchers uses machine learning to reduce solve time by up to 50 percent and produce a solution that better meets a user’s objective, such as on-time train departures. The new method could also be applied to other complex logistical problems like scheduling hospital staff or assigning airline crews.

Engineers often break down these kinds of problems into a sequence of overlapping subproblems that can each be solved in a feasible amount of time. However, the overlaps cause many decisions to be needlessly recomputed, making it take much longer to reach an optimal solution.

The researchers’ new approach learns which parts of each subproblem should remain unchanged and freeze those variables to avoid redundant computations. A traditional algorithmic solver then tackles the remaining variables.

“This is a very complex combinatorial scheduling problem,” says Cathy Wu, a member of the Laboratory for Information and Decision Systems at MIT. “Our approach can be applied without modification to all these different variants.”

The researchers’ technique, which they call learning-guided rolling horizon optimization (L-RHO), teaches a machine-learning model to predict which operations should be recomputed when the planning horizon rolls forward.

To test their approach, the researchers compared L-RHO to several base algorithmic solvers and specialized solvers. It outperformed them all, reducing solve time by 54 percent and improving solution quality by up to 21 percent.

Their method continued to outperform all baselines even when tested on more complex variants of the problem, such as factory machines breaking down or extra train congestion.

“Our approach can be applied without modification to all these different variants,” says Wu. “It even outperformed additional baselines we created to challenge our solver.”

L-RHO can also adapt if the objectives change, automatically generating a new algorithm to solve the problem – all it needs is a new training dataset.

In the future, the researchers want to better understand the logic behind their model’s decision to freeze some variables, but not others. They also want to integrate their approach into other types of complex optimization problems like inventory management or vehicle routing.

This work was supported by the National Science Foundation and MIT’s Research Support Committee, among others.

The world of professional baseball has long been aware of the importance of grip in pitching. However, a recent study by researchers from Japan has shed new light on the impact of slip between fingertips and the ball on pitching performance. Prior to June 3, 2021, Major League Baseball (MLB) pitchers had taken advantage of unapproved substances, like pine resin, to create a sticky situation that helped them maintain a precise grip. But what happens when the stickiness is removed?

A team of researchers from Tohoku University’s Graduate School of Engineering set out to understand this phenomenon. Using high-speed cameras, they captured six experienced pitchers throwing fastballs at approximately 130 kilometers per hour and analyzed how different baseball treatments impacted the finger-ball slip distance. The slip distance refers to the distance the fingers slip on the surface of the ball as it is wound and released.

The researchers found that the stickier the surface of the ball, the less the fingers slipped. This resulted in faster pitches with more revolutions per minute (RPM) and more directional control. In fact, when coated with rosin powder or pine resin, the slip distance was reduced by more than half to approximately 8 millimeters on average.

However, the study also revealed an unexpected result: when pitching water-treated balls, the velocity of the pitches dropped significantly compared to other conditions. This is thought to be due to the pitcher’s perception of fingertip slippage and subsequent adjustments in their pitching action.

The researchers’ findings are expected to enhance our understanding of the ball release mechanism under varying friction conditions, contributing to improved pitching performance, injury prevention for pitchers, and the development of better equipment.

In the future, the team plans to investigate changes in pitching movement resulting from different conditions through analysis of whole-body movements and muscle activity. They also aim to identify pitching techniques that maintain performance with slippery balls while reducing the risk of injury.

This study has significant implications for professional baseball and highlights the importance of understanding the physics behind the game. By optimizing grip and minimizing slip, pitchers can potentially improve their performance and reduce the risk of injury.

We often think of inequality as a natural byproduct of human progress, but recent research suggests that this notion may be far from the truth. A new study published in the journal PNAS has analyzed over 50,000 ancient houses from more than 1,000 sites around the world, revealing surprising patterns in house size distributions that challenge long-held views on inequality.

The researchers, led by Gary Feinman, MacArthur Curator of Mesoamerican, Central American, and East Asian Anthropology at the Field Museum in Chicago, aimed to use house sizes as a metric for wealth inequality over time. They discovered that while inequality is widespread throughout human history, it’s not inevitable, nor is it expressed to the same degree at every place and time.

“This paper is part of a larger study that looks at patterns of inequality over time,” says Feinman. “We see interpretable trends and patterns that cross-cut time and space, rather than just noise or chaos.”

The researchers compared house size distributions from different localities, finding variability in the extent of inequality across time and space. They also examined how inequality varied in relation to population, political organization, and other potential causal factors.

To quantify and compare economic inequality in different places, at different points in history, the researchers used the variable distributions of house sizes at more than 1000 settlements to calculate a Gini coefficient for each site. The coefficients were then compared across time and space to examine trends in inequality.

The results show that even while populations have risen over the years, inequality hasn’t always increased in a uniform way. “Human choice and governance and cooperation have played a role in damping down inequality at certain times and places,” says Feinman.

The study’s findings have significant implications for how we view the present and the past. While history has shown us that elements of technology and population growth can raise the potential for inequality, this potential is not always realized, as people have implemented leveling mechanisms and systems of governance that mute that potential.

Ultimately, the research suggests that inequality is not inevitable, but rather a complex outcome of human choices and institutions. As Feinman notes, “The traditional thinking expects that once you get larger societies with formal leaders, or once you have farming, inequality is going to go way up. These ideas have been held for hundreds of years, and what we find is that it’s more complicated than that.”