Field Service Routing

The world is full
of scheduling problems

For example

The world is full
of scheduling problems

Some are automated

Few are optimized

Let's optimize them
with Java and AI

Is optimization worth it?

Remember 2022?

Panic!

Reality

Quick win

• -15% natural gas consumption
• Some impact on quality of life

Imagine if we optimize

for efficiency gains without sacrifices

Vehicle routing case study

Expected: -1% driving time

Result: -25% driving time

⇒ -10 million kg CO² emission per year

⇒ -100 million $ cost per year

Devoxx scheduling

Act 1

March 2023

The bet

It's on!

Act 2

May 2023

Stephan generates a Voxxed schedule with GPT...

... and declares victory.

Feasible, not optimal

What is a feasible schedule?

No hard constraints broken.

Feasible is easy

Optimal is hard.

My mistake

There is hope

Scale from Voxxed to Devoxx

• Voxxed: 30 sessions
• Devoxx: 210 sessions
• Is Devoxx just 7 times harder than Voxxed?
• Or is it 10³⁶⁵ times harder?

Act 3

August 2023

The Devoxx BE schedule
was generated with ...

Victory!

Victory over generative AI?

Generate tags with GPT
before solving

Generative AI + Solver AI
work together

Why difficult to solve?

The web UI

Calculate score

public long calculateScore(TimeTable timeTable) {
    long hardScore = 0;
    for (Lesson lesson1 : timeTable.lessons) {
        for (Lesson lesson2 : timeTable.lessons) {
            if (lesson1.room == lesson2.room
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
            if (lesson1.teacher == lesson2.teacher
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
            if (lesson1.studentGroup == lesson2.studentGroup
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
        }
    }
    return hardScore;
}

Greedy algorithm

Greedy code

for (Lesson lesson : timeTable.lessons) {
    long bestScore = Long.MIN_VALUE;
    Room bestRoom = null;
    Timeslot bestTimeslot = null;
    for (Timeslot timeslot : timeTable.timeslots) {
        lesson.setTimeslot(timeslot);
        for (Room room : timeTable.rooms) {
            lesson.setRoom(room);
            long score = calculateScore(timeTable);
            if (score > bestScore) {
                bestScore = score;
                bestRoom = room;
                bestTimeslot = timeslot;
            }
        }
    }
    lesson.setTimeslot(bestTimeslot);
    lesson.setRoom(bestRoom);
}

DEMO

Greedy algorithm

Algorithm comparison

Brute force

Brute force code

long bestScore = Long.MIN_VALUE;
for (Timeslot timeslotLesson1 : timeslots) {
    lesson1.setTimeslot(timeslotLesson1);
    for (Room roomLesson1 : rooms) {
        lesson1.setRoom(roomLesson1);

        for (Timeslot timeslotLesson2 : timeslots) {
            lesson2.setTimeslot(timeslotLesson2);
            for (Room roomLesson2 : rooms) {
                lesson2.setRoom(roomLesson2);
                ...

                for (Timeslot timeslotLessonN : timeslots) {
                    lessonN.setTimeslot(timeslotLessonN);
                    for (Room roomLessonN : rooms) {
                        lessonN.setRoom(roomLessonN);
                        int score = calculateScore(timeTable);
                        if (score > bestScore)
                            ...

DEMO

Brute Force

How big is the search space?

Search space for n lessons: nⁿ

Search space for 400 lessons: 10¹⁰⁴⁰

Atoms in observable universe: 10⁸⁰

Algorithm comparison

Metaheuristics

• Local Search
• Tabu Search
• Simulated Annealing
• Late Acceptance
• ...

+ advanced subsystem algorithms

Implement yourself?

Planning optimization made easy

• Library of optimization algorithms
• AI, not ML
• Open Source (Apache license)
• Actively developed by our Open Core company

Easy score calculator

public HardSoftScore calculateScore(TimeTable timeTable) {
    long hardScore = 0;
    for (Lesson lesson1 : timeTable.lessons) {
        for (Lesson lesson2 : timeTable.lessons) {
            if (lesson1.room == lesson2.room
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
            if (lesson1.teacher == lesson2.teacher
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
            if (lesson1.studentGroup == lesson2.studentGroup
                    && lesson1.timeslot == lesson2.timeslot) {
                hardScore--;
            }
        }
    }
    return HardSoftScore.of(hardScore, 0);
}

DEMO

Metaheuristics

Algorithm comparison

How good are humans?

Result comparison

How good is Generative AI?

Result comparison

Constraints

Let's try this out

after the break

DEMO

Employee scheduling

How do you define constraints?

Constraint Streams

Get incremental score calculation
without writing incremental score calculation.

Why not just Java Streams?

• Incremental
• Scoring specific
• Score explanation

DEMO

Constraint Streams

Constraint Streams

• Streams-like API
• Fast and scalable
  • Incremental calculation
  • Indexed joins
• Custom code friendly

Responding to change

"Plans are of little importance,
but planning is essential."

"Everybody has a plan
until they get punched
in the mouth…"

What kind of planning disruptions?

• Local disruption
  • Impact limited to one department/region
• Global disruption
  • Impact across the board

Local disruption

Global disruptions

• 2001 - September 11 attacks
  • Air-traffic shut-down in North America for 2 days
• 2010 - Eyjafjallajökull volcano eruption
  • Air-traffic shut-down in Europe for 8 days
• 2020 - COVID-19
  • Flight cancellations

Planning agility paradox

"We can not automate planning
because the plans will change."

1) What is your
publication deadline?

Publication

• Tell employees when/how/where to work
• Publication deadline
  • Technician routing: 1 hour before departure
  • School timetabling: before end of vacation
  • Guards: 4 weeks before shift
  • Nurses: 6 weeks before shift
  • Facility location: 6 months before opening

Multi-stage planning

• Strategic: expand the maternity ward?
• Tactical: hire a respiratory specialist?
• Operational: assign a Monday shift to Ann?

Each planning stage feeds into the next.

Conway's law

"Any organization that designs a (software) system
will produce a design whose structure
is a copy of the organization's communication structure."

In practice

@Path("/trainWagon")
public class TrainWagonDeciderResource {

  @Inject
  SolverManager<TrainWagonSolution, ...> wagonSolverManager;

  ... wagonSolverManager.solve(...)
}
@Path("/trainPlatform")
public class TrainPlatformDeciderResource {

  @Inject
  SolverManager<TrainCrewSolution, ...> platformSolverManager;

  // Uses TrainWagonSolution's output as problem facts:
  // Length of train impacts eligible platforms
  ... platformSolverManager.solve(...)
}

2) Do you assign to time?

No, domain lacks time data

For example:

• Facility Location Problem
• Knapsack Problem

No java.time import

Yes, Domain has time
but planning does not assign to time

import java.time.LocalDate;

@PlanningEntity
public class Patient {

  private String name;
  private LocalDate arrivalDate;
  private LocalDate departureDate;
  ...
  @PlanningVariable(...)
  private Bed bed;
  ...
}

Hospital bed planning does not change
a patient's arrival or departure date.

By the way...

Don't use java.util.Date for time manipulation.
It's like asking your bartender how to treat cancer.

Don't use java.util.Calendar either,
It's like asking your dog how to treat cancer.

Use java.time.LocalDate or java.time.LocalDateTime instead!

Yes, Domain has time
and planning assigns to time

Pick a good domain model.

3) Do you replan
every week, day or hour?

Continuous planning

What is continuous planning?

Don't change history

@Entity
public class Shift {

  private LocalDateTime start;
  private LocalDateTime end;

  @PlanningPin
  private boolean history;
  ...
  @PlanningVariable(...)
  private Employee employee;
  ...
}

Don't change published shifts?

@PlanningEntity
public class Shift {

  private LocalDateTime start;
  private LocalDateTime end;

  @PlanningPin
  // every historic shift is published
  private boolean published;
  ...
  @PlanningVariable(...)
  private Employee employee;
  ...
}

Maybe

How to paint yourself in a corner

Assign all respiratory specialists
to the last shift of your planning window
(even in other wards as normal nurses).

Continuous planning

• Publish notice ranges from long to short
  • Weeks: employee shift rostering, pharmacy on duty scheduling, ...
  • Seconds: runway scheduling, gate scheduling, platform scheduling, ...
• Domain must contain dates or timestamps data
  • So domain probably has import java.time
  • Regardless if planning assigns to time or not.
• Planning windows must overlap
  • ⇒ School timetabling is not continuous planning

4) Will published plan change?

Non-disruptive Replanning

Implementation

@PlanningEntity
public class Talk {

    private Timeslot publishedTimeslot;
    ...

    @PlanningVariable(...)
    private Timeslot timeslot;

}

  Constraint publishedTimeslot(ConstraintFactory f) {
    return f.forEach(Talk.class)
        .filter(talk -> talk.getPublishedTimeslot() != null
            && talk.getTimeslot()
                != talk.getPublishedTimeslot())
        .penalize(HardMediumSoft.ONE_MEDIUM)
                .asConstraint(PUBLISHED_TIMESLOT);
  }

Non-disruptive Replanning

• Medium penalty: disrupt only to become feasible
  • Timeslot in conference scheduling
• Soft penalty: disrupt if gain is higher than a threshold
  • Room in conference scheduling

5) Does your problem
change every few (milli)seconds?

Real-time planning

Real-time planning

DEMO

Real-time planning

• solver.addProblemFactChange(...)
• Daemon mode
  • Don't waste CPU time at night and during breaks

7) Do you have
a resource shortage?

Overconstrained planning

Overconstrained planning

• Nullable
  • @PlanningVariable(nullable = true)
  • medium constraint penalizes unassigned entities
  • Not yet supported for VRP, use dummy vehicle workaround.
• Virtual resources
  • Add virtual resources to the input problem

    class Vehicle {
      boolean virtual;
    }

  • Soft constraint penalizes cost of virtual resources

Benchmarker

Benchmarker

DEMO

Get started

Start coding today

1. Go to timefold.ai
2. Click the Quickstarts repo button.
3. Pick a quickstart from the README.
4. Run it:

$ git clone https://github.com/TimefoldAI/timefold-quickstarts.git
...
$ cd timefold-quickstarts/use-cases/vehicle-routing-time-windows
$ mvn quarkus:dev
...

Q & A

Learn more	timefold.ai
Feedback	@GeoffreyDeSmet @GeoffreyDeSmet@mastodon.social
Get started