Tag: scheduling

OM in the News: Delta’s Unorthodox Scheduling System

Delta's Control Room
Delta’s Control Room

“The crew of Delta Air Lines  Flight 55 last Thursday couldn’t legally fly from Lagos, Nigeria, to Atlanta unless they waited a day due to new limits on how much pilots can fly in a rolling 28-day period,” writes The Wall Street Journal (April 3, 2014). The trip would have to be canceled. Instead, Delta headquarters told the captain to fly to San Juan, which they could reach within their duty limits. There, two new pilots would be waiting to take the Boeing 767 on to Atlanta. The plane arrived in San Juan at 2:44 a.m., quickly took on fuel and pilots, and landed in Atlanta only 40 minutes late.

The episode, unorthodox in the airline industry, illustrates the fanaticism Delta now has for avoiding cancellations. Last year, Delta canceled just 0.3% of its flights. That was twice as good as the next-best airlines, Southwest and Alaska, and five times better than the industry average of 1.7%.

As it cut cancellations with a more-reliable operation, overall on-time arrivals improved and Delta has fewer delays. Managers in Delta operations center (featured in our Global Company Profile  in Chapter 15) move planes, crews and parts around hourly trying to avoid canceling flights. How well an airline maintains its fleet and how smartly it stashes spare parts and planes at airports affect whether a flight goes or not. Delta’s new analytical software and instruments that can help monitor the health of airplanes and predict which parts will soon fail. Empty planes are ferried to replace crippled jets rather than waiting for overnight repairs. Typically the airline has about 20 spare airplanes of different sizes each day. About half are stationed in Atlanta and the rest spread around other domestic hubs and two in Tokyo.

Classroom discussion questions:

1. Why have Delta’s operations managers focused on cancelling fewer flights?

2. How does Delta’s fleet age (one of the oldest in the industry) impact this strategy?

OM in the News: New Looks at Loading and Boarding Planes

alaska air jetwayLoading an airplane quickly and efficiently isn’t an easy task. “It should be, and could be, but the humans involved can’t seem to get with the program,” writes Wired Magazine (Aug. 28, 2013). Better loading means more time in the air – which is where airlines make their money. The boarding process is far from standard – there are almost as many boarding procedures as there are airlines. This problem has long been pondered by operations managers, without a definitive answer. But there have been a few promising experiments.

The most unusual and deceptively simple idea is opening the door at the rear of the plane in addition to the door at the front. Alaska Airlines is trying this. The idea isn’t entirely new–many airlines open the front and rear doors at those airports where there is no jetway, only a staircase leading to the tarmac. Alaska has a new tool to help facilitate using both doors–a solar-powered ramp. Mounted on wheels, the ramp can be driven to the backdoor of the airplane, and passengers make two switch-back turns down the ramp to the ground, providing an alternative to stairs for easy suitcase rolling and wheelchair access. Using the aft door to unload passengers can reduce the turnaround time by 10 minutes.

One of the big reasons boarding has slowed to a crawl is people are carrying more bags aboard to avoid baggage fees. So American Airlines is experimenting with letting those who checked their bag board first. Ideally, these passengers will simply walk to their row and sit down. The airline says that overall it has shaved a few minutes off the boarding process.

Although airlines commonly board by sections, it’s generally a free-for-all with regard to where in that section you are. United uses the “outside-in” method of seating window passengers first, then middle, then aisle seats. The airline has been organizing passengers in better defined lines at the gate for each group, with the hope is there will be less of a bottleneck.

Discussion questions:

1. Why is boarding an important OM issue?

2. What else can airlines do to speed the process?

OM in the News: Speeding Up the Airline Boarding Process

airline arrival rates“The forecast calls for heavy frustration with a 50% chance of innovation at U.S. airports this summer,” writes The Wall Street Journal (May 23,2013). Airport crowds are expected to be the largest in the U.S. since 2008. There’s already concern that budget cuts in TSA overtime will lead to longer security-screening lines. But travelers will find new boarding procedures at both United and American airlines that pack planes faster. The idea is to increase the percent of flights that arrive within 15 minutes of schedule, currently around 76%.

In an effort to get more flying out of planes and crews, a topic we discuss in Chapter 15, United is introducing new boarding lanes at gate areas this summer. Five different boarding groups will line up in different areas, akin to how Southwest lines up customers by groups, so that instead of a crush of people pushing toward the gate, each group will have a designated place to wait. After elite-level customers, the rest of the coach cabin will board window-seat passengers first, then middle seats, and aisle seats last. With the “Wilma” system, as United calls it, seats fill faster because people already seated don’t have to get up as much to let a row mate in. United says Wilma boarding is about 20% faster than boarding from the rear of the plane to the front.

American set its own new change in boarding this month. It joined Alaska Airlines in offering early boarding to customers who don’t have large carry-on bags. American offers Group 2 boarding, right after elite-level frequent fliers, to passengers without overhead-bin luggage. American tested the idea in seven cities earlier this year and found getting people without overhead bin luggage on early sped up boarding time and improved on-time arrival performance.

Discussion questions:

1. Why has Southwest’s management of scheduling been so successful?

2. What are the advantages and disadvantages of the “Wilma” system?

Guest Post: Tying Scheduling Back to Operations Strategy in Your Course

steve harrodDr. Steven Harrod is Assistant Professor of Operations Management at the University of Dayton and can be reached at steven.harrod@udayton.edu.

For those of you including Chapter 15, Short-Term Scheduling, in your syllabus, here is a great way to tie that topic back to the beginning of your course and to Chapter 2, Operations Strategy in a Global Environment. A significant topic in Chapter 15 is “Sequencing Jobs”, which is a logical extension to Module D– Waiting-Line Models.

In class, I work out three fundamental queue disciplines: first come first serve (FCFS), shortest processing time (SPT), and earliest due date (EDD). Following Example 5 from Chapter 15 of the Heizer/Render text, I walk my class through the solution of these three sequences using a custom worksheet that you may obtain online by clicking here.

I ask the students to recall the three competitive advantages of Chapter 2 (low-cost, differentiation, and response), and how sometimes it is unclear which competitive advantage applies. For example, does McDonald’s pursue a low-cost or differentiation strategy?

I then ask the students to label each sequencing rule by the competitive advantage it best aligns with. Clearly, EDD best supports response, because it most respects the timeliness of delivery. Then I impress upon the students how SPT leads to low-cost competitive advantage (WIP, flowtime). Finally, I assert that FCFS is a differentiation strategy, because it enforces social justice in the service pattern, and thus seeks to make each customer feel valued and unique. Viewed from the way in which they queue their customers, it then becomes obvious that McDonald’s pursues a low-cost strategy, Wendy’s pursues a differentiation strategy, and Domino’s pursues a response strategy.

I close this lecture by impressing on students how choices in Operations Management effectively dictate the strategic competitive advantage of the firm. Firms must align their stated strategy and their operating rules to each other, or else be perpetually in a mode of crisis and confusion.

OM in the News: Art and Science of Scheduling the N.F.L.

Howard Katz, NFL scheduling tzar

“We’re geniuses one day and absolute morons the next,” says Howard Katz, director of scheduling for the  National Football League. That’s because Katz must consider a confounding array of factors, from the N.F.L.’s expanded Thursday night package, which gives each team a game in a short week, to potential baseball playoff situations that could affect the availability of stadiums and parking lots in October.

The New York Times (April 20,2012) reports that for the networks that pay billions of dollars to carry N.F.L. games, Katz’s staff has been mostly geniuses. N.F.L. games were watched by an average of 17.5 million viewers last season. N.F.L. games accounted for 23 of the 25 most-watched television shows among all programming, and the 16 most-watched shows on cable last fall.

Designing a schedule that generates those ratings, while also guaranteeing competitive fairness, is more complicated than ever, even though software spits out 400,000 complete or partial schedules (once done entirely by hand) from a possible 824 trillion game combinations. Katz starts with thousands of seed schedules, empty slates in which a handful of critical games with attractive story lines are placed in select spots. Then the computers generate possibilities around those games.

The N.F.L. also feeds the computer with penalties for situations it prefers to avoid — three-game trips, for example, or teams starting with two road games. There are requests not to play at home on certain holidays — the Jets and the Giants typically ask not to play home games during the Jewish High Holy Days.  This year, the software generated 14,000 playable schedules, which were reduced to 150 with an eyeball test. Katz reviewed those 150 by hand, scoring them for each team and each network.

Linear programming may be at the heart of scheduling, but the process is definitely part art and part science.

Discussion questions:

1. Why is scheduling sports teams so complex?

2. Are all the teams happy with the final schedules?

Good OM Reading: Using LP to Schedule NCAA Basketball Tournament Games

What could be more timely than an article in the Journal of the Operations Research Society called “Team Assignments and Scheduling for the NCAA Basketball Tournament.”  The paper, by U. of Alabama professors S.H. Melouk and B.B. Keskin, provides a wonderful example to use in class when you teach linear programming, in Module B.

The authors write: “The buzz of the tournament and the fanatical behavior of the followers of the participating teams serve as our motivation to examine and develop a team assignment model that maintains the integrity of the tournament while also attempting to place teams closer to their campus location, thus making it easier for both fans and teams to travel to the game sites. Observation of game venues shows a decrease in the actual attendance at early round tournament games. In 2010, actual attendance at the early round game sites was, on average, 83.5% of capacity. This statistic is surprisingly low. A likely contributing factor is the long distances that fans must travel to attend games.”

The growing NCAA concern is travel expenses of the participating teams, as the NCAA reimburses each team for their travel to tournament games. Given there are 68 tournament teams, it is a significant expense to transport the players, their equipment, and coaching staffs to game sites. In an effort to curb expenses, the NCAA  requires a minimum distance of 350 miles from a game site before air travel is reimbursable.

The article describes the development of an integer LP program designed to optimize team assignments in the sense of minimizing the total distance travelled by teams to game sites. Results of testing the model against actual tournament assignments  show consistent and significant cost savings and reductions in distance travelled. In fact, 28,202 travel miles were saved in 2010 with use of the LP model.

 

OM in the News: Is Boarding the Most Annoying Airline Delay?

I guess the answer is yes, given that yesterday’s New York Times (Nov. 1, 2011) ran a front page article on the woes of airline boarding.  Airlines have been complicating the process of boarding  for decades, writes the Times. First travelers have to be sorted by priority: 1st class, frequent flyers, elite cardholders, military in uniform, families with kids, and so on. Then there is the issue of seats  for those who paid priority boarding fees, the matter of more roll-ons in the aisle from checked-baggage fees, and of course,  the fact that planes are fuller. Boarding time has actually doubled— and it now takes 30-40 minutes to board 140 passengers, up from 15 minutes about 20 years ago.

 Operations scheduling techniques (Ch.15) have been used by every airline to find creative solutions to speed boarding. Spirit Airlines says the answer is to charge $20-$40 per carry-on bag. This “stress-free” boarding saves 6 minutes on average. American Airlines switched a few months ago to boarding passengers earlier who pay $9-$19 extra (and hence find a space for their bag). The rest of the passengers are brought in as 3 groups, sorted out in a spread throughout the plane. This method has cut boarding by 4-5 minutes.

US Airways hired ASU profs to develop a “reverse pyramid” to save time. Passengers with window seats in the back board first. Then, gradually, passengers are brought on to the front of the plane in a staggered pattern. Southwest, which can board its planes in 15 minutes, claims the root of delays is the practice of assigning seat numbers. So it assigns passengers to one of 3 boarding groups.

The conflict: all the extra fees for early boarding and baggage will add $12.5 billion to the bottom  line of US airlines this  year, up 87%  from last year.

Discussion questions:

1. Research has shown that “back-to-front” boarding is the slowest method. Why is it still used?

2. How else can OM help solve this problem?

OM in the News: Scheduling Major League Baseball Umpires

Here is an interesting scheduling application you may want to share with your class when you teach Chapter 15. Scientific American (Aug.18,2011) reports on how four B-School profs have formulated the “travelling umpire problem” to develop solutions to get umpire crews to every major league baseball game . Given that the Major League baseball (MLB) season lasts 6 months, such scheduling is a daunting task.  During the season, 30 teams play a total of 2,430 games in 27 different cities. The umpires in the league are part of a 4-member group called a crew and each umpire handles about 142 games/year.

Here are some constraints: (1) minimize travel time and distance for the crews; (2) crews should visit each MLB city at least once; (3) they should work each team at home and on the road; (4) they should work no more than 21 days in a row; (5) they should not ump any one team’s games for more than 4 series all year–just to name a few of the rules.

The mathematical model proved successful in generating a high quality schedule in a short amount of time and MLB has used it over the past 3 seasons. Before the profs (who are at U. Miami, Carnegie, and Michigan State) built their computerized method, the schedule was created manually–and took weeks– by a retired umpire. As Scientific American puts it: “That guy is out“!

Researcher Tallys Yunes (at Miami) explains, “We not only reduced the time necessary to create the schedule, we also improved the overall quality of the schedule, in the sense that it better satisfies both the MLB and umpire union rules”.

If you want to provide a humourous side-bar to this class discussion here is a link to a 4- minute video clip about umpiring. It features Leslie Nielson playing a detective going undercover as an ump.

Discussion questions:

1. Why is this an important OM issue?

2. Besides major league sports, what other fields could benefit from math scheduling models like this?

OM in the News: A Pill For Night-Shift Workers?

A few months back, my friend CJ took a 9-5 job in a factory and was happy to be back in the ranks of the employed. Shortly after he started, he was told he needed to also pick up a night shift (2-11 am) after his regular shift once a week. We discuss the problems of scheduling workers for varying shifts in Chapter 15, where we note that falling asleep on the job is not at all rare. Sure enough, CJ’s body never could adjust to the constant changes. Coffee used to work for me when I drove double shifts of busses for the CTA in Chicago while in college. But now Businessweek (Aug.11-18, 2011) reports a potential new cure for the 15 million Americans facing “shift work sleep disorder”.

It comes in the form of a new pill made by Cephalon, called Nuvigil, an upgraded version of the blockbuster narcolepsy drug Provigil. With Provigil’s patent protection running out next year, Cephalon is spending big to promote the “new” product. Some doctors, though, worry about a drug solution for staying awake on the job, especially because the pill can be both addictive and carries  side effects like nausea, skin rashes, hallucinations, and depression. “Caffeine is a very good wake-promoting agent, and it’s a lot cheaper”, says a Columbia U. doctor. Adds a Cleveland Clinic physician: “We want to treat the real condition, rather than just papering over the symptoms with a medication that can just keep people awake longer”.

Still, Cephalon’s campaign is working. With the malady affecting one in 4 shift workers, Nuvigil sales are  growing 50% annually for the $12 pill, enough for Israel’s Teva Pharmaceutical to agree to buy Cephalon for $6.2 billion.

Discussion questions:

1. Should employers consider providing Nuvigil to their shift workers?

2. What are the ethical implications of producing such a drug?

OM in the News: Running an Airline with No Scheduled Flights

Scheduling crews, planes, flights, and fuel at Delta (722 aircraft), United (710), and American (618) is a topic we discuss in Ch. 15, as well as in Module B (LP) in our texts. But what about OM  and scheduling at the nation’s 4th largest domestic air carrier–NetJets– with over 600 planes, 2,500 pilots, and 270 flight attendants?  Air & Space Magazine (Aug.1, 2011) reports today how NetJets produces a flight schedule based on the unpredictable travel needs of its 7,000+ well-heeled fractional jet owner customers. 

“An insane ballet that requires constant attention to detail”, is how a NetJets manager describes a flight schedule that will likely be reduced to little more than junk before the sun sets. Whereas ordinary airlines operate flight schedules that change little from day-to-day, at airlines on demand, no two days are the same.  This ” is one of aviation’s most logistically complex endeavors”, company execs assert.

In the early days, NetJets tracked aircraft on metal boards with a 24-hour grid. Today, a proprietary software program, maintained by 200 IT employees, factors in every conceivable variable– from plane weight to weather to sunset restrictions—at every airport. It makes sure the client’s limo and hotel are coordinated and will downsize a plane if a runway is slippery.

But no amount of software can account for curveballs the customers throw. “I think I might want to leave at 8, but it could be 10. I may have 4 people, but it could be 8. I may have a dog but I may not. By the way, I don’t want to go from Palm Beach to NY. I want to go to Moscow”, are just some of the  calls to customer service reps. ” We don’t say no–ever”, says the VP-Scheduling. An aircraft is guaranteed 4-10 hours after a request is made. The crew work schedule: on 1 week, off 1 week, with no idea where they are flying till minutes before takeoff.

Discussion questions:

1. Compare OM at NetJets to a traditional airline.

2. What is the most difficult aspect of operating an airline with no fixed schedule?

OM in the News: How to Get Passengers to Board the Plane Faster

In Chapter 15, we discuss Southwest’s strategic advantage in being able to turn around its planes faster than competitors. Part of that process means getting passengers on board and ready to go as quickly as possible. Yet, as The Wall Street Journal (July 21,2011) points out, “Boarding an airline can be a bit like the after-Christmas sale at Wal-Mart.  Passengers jockey to get better positions in line. The aisles become clogged with travellers stuffing luggage the size of a 4th-grader into overhead bins”. To address the problem at American, that airline just finished a 2-year study to try to speed up its boarding.

The result: AA rolled out a new strategy–randomized boarding. Travellers without elite status now get assigned randomly to boarding groups instead of filing onto planes from back to front. American says the new system can shave 3-4 minutes off the average 20-25 minutes. And every minute cut saves the airline $30/flight.

After first observing 1,000’s of arrivals and departures to see where the process slowed down, American found that one time factor was baggage–more bags are being carried on to avoid fees. “Back-to-front” slowed because only 2 people on average got to their seats at a time, while everyone else standing and waiting filled bins at the front of the plane —and was the most time-consuming. ( Alaska Airlines, US Airways and Continental all use “back-to-front”, by the way).

Computer simulations revealed that “window-middle-aisle” (used by United Airlines and Delta) –meaning boarding passengers in window seats 1st, followed by middle  and then aisle–was faster.  But randomized boarding worked even better. Multiple passengers got to their seats at the same time. Bins filled more evenly. The process reduced the number of bags that needed to be checked at the gate by 20% because more overhead space was available. And, the system proved calmer when tested on real flights.

Discussion questions

1. Why do the airlines use such diverse boarding systems?

2. Why does Southwest, which uses a “1st-to-check-in” boarding system, turn its planes around faster than other airlines?

OM in the News: From Lady Gaga to the Orlando Magic

What do Lady Gaga and the Orlando Magic have in common? As the Orlando Sentinel (April 15,2011) reports, Lady Gaga (if you don’t know who she is, don’t worry–your students do) performed her Monster Ball Tour on April 15th in the sparkling new Amway Center in downtown Orlando before 17,000 screaming fans. It took 42 trucks and busses to carry her massive concert set to town. The unloading and set building process began the night before and was only possible because the Amway Center has state-of-the-art technology and space for the show. The old Amway Arena had only half the number of loading bays, no freight elevators, and less sophisticated sound and lighting–hence missed out as a destination for major tours.

How is this an operations management issue? It’s because today (April 16th), the Orlando Magic open their first NBA Playoff game of the season against the Atlanta Hawks in the very same space. With ESPN providing national TV coverage in 3-D, Lady Gaga’s crew had just 3 hours to clear out-of-town (from 11:30pm yesterday till 2:30am today) to make room for the ESPN crew to set up, and to allow for the concert floor and seats to be converted to a basketball arena.

Precise scheduling,  excellent processes, and good communications make these change-overs possible. Whether its Lady Gaga to Magic, or NY Knicks to Ice Follies, OM is the key to success.

Discussion questions:

1. Name some other activities that require such perfect scheduling.

2.  Why are ultra-modern arenas so important to cities like Orlando?

OM in the News: Sleeping Air Traffic Controllers

In Chapter 15 (Short-Term Scheduling), we discuss what Three Mile Island and Chernobyl’s nuclear power plants had in common: drowsy “graveyard shift” workers trying to stay awake amid constantly changing work schedules. The same, of course, is true for pilots who often fall asleep in the cockpit on long hauls (like Atlanta-to-Mumbai’s 18-hour flight), just as it is for truck drivers, assembly line workers, and virtually anyone who is a sleep-deprived shift worker. Scheduling is a major problem in all firms with 24/7 shifts.

So when The New York Times and The Wall Street Journal (April 14,2011) both reported that the lone air traffic controller at Reno Tahoe  Airport  fell asleep on the job early yesterday (and didn’t respond to pilot calls for 16 min.), it’s hard to be shocked. (Of course, we hope the situation is rectified before the POMS Meeting in Reno in a few weeks).

It was the 5th time in the past 3 weeks in which controllers were found to be sleeping. The most memorable one was at Washington’s Reagan National Airport on March 23rd, at midnight, forcing pilots of 2 passenger jets to land on their own. The recent incidents are just “the tip of the iceberg”, says a Boeing safety exec.

 “We absolutely cannot and will not tolerate sleeping on the job”, says the FAA administrator. The solution:  a 2nd employee will be added to the late shift at every airport that operates overnight. But inconsistent work schedules contribute to fatigue, and the FAA is for the 1st time, trying to fashion a program to allow controllers to take naps.

Discussion questions:

1. Have any of your students worked late shifts and what is their opinion of the problem/solution?

2. How do pilots  and airlines handle the fatigue factor?

Video Tip: Scheduling Employees at Hard Rock Cafe

One of my favorite videos is a short  (4.5 min.) study of how Hard Rock Cafe schedules its 160 servers at the giant 1,100 seat Hard Rock here in Orlando. I show it when I teach scheduling (Ch.15) and linear programming (Mod.B in the hard cover text). The  topic is one  many students relate to, especially if they have worked in retail or restaurants, where schedules are always a sensitive subject.

In Hard Rock’s case, the sales forecast is critical. Many factors are considered in deciding how many servers to call in, including historical sales, major conferences in town, season, etc. Each employee submits a weekly request form, and then an LP package takes over, with the objective of minimizing the number of employees per shift. It turns out that the system works quite well and employees are usually satisfied. Turnover, even during non-recession times, is 1/2 the industry average.

What we don’t mention in the video is that the managers never mastered the scheduling software, which is actually somewhat complex. But one, very enterprising, young server offered to handle the weekly task on his own. He collects all the forms, goes down into a basement office every Saturday, where it takes him about 6 hours to input the data and churn out the schedules. He does this for no additional pay! Why, you ask? Because  constraints and schedules are set by seniority, and he is allowed to assign himself the highest priority, a 9. It turns out that a great schedule, at the right work stations, can make the difference of $100’s a week in tips.

This topic is one that students with jobs are more than happy to discuss..