Reasoning with Conditional Time-intervals

Reasoning with
Conditional Time-intervals
P. Laborie and J. Rogerie
[plaborie,jrogerie]@ilog.fr

FLAIRS 2008 – Coconut Grove, Florida, USA – May 15-17, 2008
Overview
 Motivations
 Model
 Constraint Propagation
 Extensions & Conclusion

Motivations
 Optional time-intervals in scheduling:
 Some activities can be left unperformed
A C
B

Motivations
 Some activities can be left unperformed
A
B

Motivations
 Some parts of the schedule can be left unperformed
A
B
C
D
E

Motivations
A
B

C
D
E
Motivations
A
B
 Dependencies: exec(E)exec(C)exec(D)

C: Mode 1
Motivations
 Alternative combination of resources
A
B
C
XOR
C: Mode 2
C: Mode 3

Motivations
A
B
CC:: MMooddee 1

Motivations
A
B
CC:: MMooddee 22

Motivations
A
B
CC:: MMooddee 33

Motivations
 Alternative recipes (common in manufacturing)
A
B
XOR
C
C: Recipe 1
Op11 Op12 Op13
C: Recipe 2
Op23
Op21 Op22

Motivations
A
B
Recipe C: recipe 1
Op11 Op12 Op13

Motivations
A
B
C: Recipe 2
Op23
Op21 Op22

Motivations
 Objective: Make it easier to model and solve
scheduling problems involving optional
activities/tasks/processes/recipes/…
 Constraint Optimization framework
 Basic ingredients:
 Optional time-intervals variables (a,b,c,… )
 Logical constraints (exec(a)  exec(b))
 Precedence constraints (endBeforeStart(a,b))
 Decomposition constraints (span(a,{b1,…bn}))
 Alternative constraints (alternative(a,{b1,…bn}))

Focus of this talk
MODEL PROBLEM SOLVING
Basic constraints:
logical, precedence,
span, alternative Basic constraint
Time-interval
variables
propagation

What we won’t talk about here
MODEL PROBLEM SOLVING
Resource-related
Basic constraints:
logical, precedence,
span, alternative Basic constraint
Time-interval
variables
propagation
constraints:
sequencing, cumul,
states, calendars
Resource-related
constraint
propagation
Search &
optimization
techniques

Model: Time interval variables
 Extension of classical CSPs
 A new type of first class citizen decision
variable: Time-interval variable
 Domain of values for a time-interval variable a :
Dom(a)  {}  { [s,e) | s,eℤ, s≤e }
Non executed Interval of integers

Model: Time interval variables
 Domain of values for a time-interval variable a :
Dom(a)  {}  { [s,e) | s,eℤ, s≤e }
 Notations: Let a be a fixed time-interval variable
 If a={[s,e)} (a is executed), we denote:
x(a)=1, s(a)=s, e(a)=e, d(a)=e-s
 If a={} (a is non-executed), we denote:
x(a)=0 (in this case, s(a),e(a),d(a) are meaningless)

Model: Logical constraints
 Execution unary constraint exec(a) means that a
is executed (x(a)=1)
 2-SAT clauses over execution constraints:
[¬]exec(a)  [¬]exec(b)
 Expressivity:
 Same execution status:
¬exec(a)exec(b),exec(a)¬exec(b)
 Incompatibility: ¬exec(a)¬exec(b)
 Implication: ¬exec(a)exec(b)

Model: Precedence constraints
 Simple Precedence Constraints ti+z≤tj reified by
execution statuses
 Example: endBeforeStart(a,b,z) means
x(a)x(b)  e(a)+z≤s(b)
 startBeforeStart, startBeforeEnd
endBeforeStart, endBeforeEnd
startAtStart, startAtEnd
endAtStart, endAtEnd

Model: Span constraint
 Span constraint span(a,{b1,…,bn}) means that if a
is executed, it spans all executed intervals from
{b1,…,bn}. Interval a is not executed iff none of
intervals {b1,…,bn} is executed.
b1 b4
b3
b5
a
b2

Model: Alternative constraint
 Alternative constraint alternative(a,{b1,…,bn})
means that if a is executed, then exactly one of
the {b1,…,bn} is executed and synchronized with
a. Interval a is not executed iff none of intervals
{b1,…,bn} is executed.
a
b1
b2
b3
XOR

Model: Simple example
 Inspired from [Barták&Čepek 2007]
CollectMaterial (1)
GetTube
MakeTube
SawTube (30) XOR
ClearTube (20)
SawRod (10)
ClearRod (2)
WeldRod (15)
BuyTube (40)
WeldTube (15)
AssemblePiston (5)
ShipPiston (0)
CollectKit (1)
AssembleKit (5)
Deadline=70

Model: Simple example (ILOG OPL Studio)

Constraint Propagation: Interval variables
 Time interval variable domain representation:
tuple of ranges:
 [xmin,xmax][0,1]: current execution status
 [smin,smax] ℤ: conditional domain of start time would
the time-interval be executed
 [emin,emax] ℤ: conditional domain of end time would
 [dmin,dmax] ℤ+: conditional domain of duration would

Constraint Propagation: Logical network
 Logical constraints are aggregated in an
implication graph: all 2-SAT logical constraints
[¬]exec(a)  [¬]exec(b) are translated as
implications ( ¬[¬]exec(a)  [¬]exec(b) )
 Incremental transitive closure of the
implication graph allows detecting infeasibilities
and querying in O(1) whether exec(a)  exec(b)
for any (a,b)

Constraint Propagation: Temporal network
 Precedence constraints are aggregated in a temporal
network

network
 Conditional reasoning:
From logical network
exec(a)exec(b)
a b endBeforeStart(a,b)

network
exec(a)exec(b)
 Propagation on the conditional bounds of a (would a be
executed) can assume that b will be executed too, thus:
emax(a)  min(emax(a), smax(b))

network
exec(a)exec(b)
 Propagation on the conditional bounds of a (would a be
executed) can assume that b will be executed too, thus:
emax(a)  min(emax(a), smax(b))
 Bounds are propagated even on time intervals with
still undecided execution status !

Model: Simple example
 Inspired from [Barták&Čepek 2007]
CollectMaterial (1)
GetTube
BBuuyyTTuubbee ((4400))
MakeTube
SawTube (30) XOR
ClearTube (20)
SawRod (10)
ClearRod (2)
WeldRod (15)
BuyTube (40)
WeldTube (15)
AssemblePiston (5)
ShipPiston (0)
CollectKit (1)
AssembleKit (5)
Deadline=70

Extensions & Conclusion
 Conditional time-intervals are the foundation of
the new version of ILOG CP Optimizer (2.0) for
modeling and solving detailed scheduling
problems
 Model extensions:
 Resources (sequencing, cumulative, state reasoning)
 Calendars (resource efficiency curves, days off, etc.)
 Expressions to use interval bounds (start, end, etc.) in
classical CSP constraints on integer variables

Extensions & Conclusion
 Conditional time-intervals are the foundation of
the new version of ILOG CP Optimizer (2.0) for
modeling and solving detailed scheduling
problems
 Search:
 Extension of SA-LNS [Laborie&Godard 2007] to handle
optional time-interval variables

Example: Multi-Mode RCPSP

Data

Model

Trial Version:
http://ilog.com/products/oplstudio/trial.cfm

Reasoning with Conditional Time-intervals

More Related Content

More from Philippe Laborie (19)

Recently uploaded (20)

Reasoning with Conditional Time-intervals