Philippe Laborie | IBM Research (original) (raw)

Papers by Philippe Laborie

HAL (Le Centre pour la Communication Scientifique Directe), 1997

International audienc

Proceedings of the International Conference on Automated Planning and Scheduling, Jun 5, 2017

A challenging Earth-observing satellite scheduling problem was recently studied in (Frank, Do and... more A challenging Earth-observing satellite scheduling problem was recently studied in (Frank, Do and Tran 2016) for which the best resolution approach so far on the proposed benchmark is a time-indexed Mixed Integer Linear Program (MILP) formulation. This MILP formulation produces feasible solutions but is not able to prove optimality or to provide tight optimality gaps, making it difficult to assess the quality of existing solutions. In this paper, we first introduce an alternative disjunctive MILP formulation that manages to close more than half of the instances of the benchmark. This MILP formulation is then relaxed to provide good bounds on optimal values for the unsolved instances. We then propose a CP Optimizer model that consistently outperforms the original time-indexed MILP formulation, reducing the optimality gap by more than 4 times. This Constraint Programming (CP) formulation is very concise: we give its complete OPL implementation in the paper. Some improvements of this CP model are reported resulting in an approach that produces optimal or near-optimal solutions (optimality gap smaller than 1%) for about 80% of the instances. Unlike the MILP formulations, it is able to quickly produce good quality schedules and it is expected to be flexible enough to handle the changing requirements of the application.

Lecture Notes in Computer Science, 2018

We consider a well known resource allocation and scheduling problem for which different approache... more We consider a well known resource allocation and scheduling problem for which different approaches like mixed-integer programming (MIP), constraint programming (CP), constraint integer programming (CIP), logic-based Benders decompositions (LBBD) and SAT-modulo theories (SMT) have been proposed and experimentally compared in the last decade. Thanks to the recent improvements in CP Optimizer, a commercial CP solver for solving generic scheduling problems, we show that a standalone tiny CP model can out-perform all previous approaches and close all the 335 instances of the benchmark. The article explains which components of the automatic search of CP Optimizer are responsible for this success. We finally propose an extension of the original benchmark with larger and more challenging instances.

HAL (Le Centre pour la Communication Scientifique Directe), Sep 8, 2014

ABSTRACT The Time-Dependent Traveling Salesman Problem (TDTSP) is the extended version of the TSP... more ABSTRACT The Time-Dependent Traveling Salesman Problem (TDTSP) is the extended version of the TSP where arc costs depend on the time when the arc is traveled. When we consider urban deliveries, travel times vary considerably during the day and optimizing a delivery tour comes down to solving an instance of the TDTSP. In this paper we propose a set of benchmarks for the TDTSP based on real traffic data and show the interest of handling time dependency in the problem. We then present a new global constraint (an extension of no-overlap) that integrates time-dependent transition times and show that this new constraint outperforms the classical CP approach.

Computers & Operations Research, Nov 1, 2020

In this work, the online printing shop scheduling problem is considered. This challenging real pr... more In this work, the online printing shop scheduling problem is considered. This challenging real problem, that appears in the nowadays printing industry, can be seen as a flexible job shop scheduling problem with sequence flexibility in which precedence constraints among operations of a job are given by an arbitrary directed acyclic graph. In addition, several complicating particularities such as periods of unavailability of the machines, resumable operations, sequence-dependent setup times, partial overlapping among operations with precedence constraints, release times, and fixed operations are also present in the addressed problem. In the present work, mixed integer linear programming and constraint programming models for the minimization of the makespan are presented. Modeling the problem is twofold. On the one hand, the problem is precisely defined. On the other hand, the capabilities and limitations of a commercial software for solving the models are analyzed. Extensive numerical experiments with small-, medium-, and large-sized instances are presented. Numerical experiments show that the commercial solver is able to optimally solve only a fraction of the small-sized instances when considering the mixed integer linear programming model; while all small-sized and a fraction of the medium-sized instances are optimally solved when considering the constraint programming formulation of the problem. Moreover, the commercial solver is able to deliver feasible solutions for the large-sized instances that are of the size of the instances that appear in practice.

SmartDeliveries is a city-wide delivery rounds optimization system aiming at leveraging city traf... more SmartDeliveries is a city-wide delivery rounds optimization system aiming at leveraging city traffic information to optimise professional vehicle rounds. This system has been developed as part of the OptimodLyon project. One of the key features of the system is to integrate information of short and longer term expected mobility demand. In this paper, we provide a detailed presentation of SmartDeliveries system and some evaluation results that show its potential to both optimize urban deliveries (18% savings in distance and 11% in time) and, if adopted at the scale of a city, significantly contribute to globally improved traffic (5% reduction in traffic and corresponding reduction in emissions).

Lecture Notes in Computer Science, 1999

Lecture Notes in Computer Science, 2018

This paper presents the concept of objective landscape in the context of Constraint Programming. ... more This paper presents the concept of objective landscape in the context of Constraint Programming. An objective landscape is a light-weight structure providing some information on the relation between decision variables and objective values, that can be quickly computed once and for all at the beginning of the resolution and is used to guide the search. It is particularly useful on decision variables with large domains and with a continuous semantics, which is typically the case for time or resource quantity variables in scheduling problems. This concept was recently implemented in the automatic search of CP Optimizer and resulted in an average speed-up of about 50% on scheduling problems with up to almost 2 orders of magnitude for some applications.

International Joint Conference on Artificial Intelligence, 2005

This paper describes a simple complete search for cumulative scheduling based on the detection an... more This paper describes a simple complete search for cumulative scheduling based on the detection and resolution of minimal critical sets (MCS). The heuristic for selecting MCSs relies on an estimation of the related reduction of the search space. An extension of the search procedure using selfadapting shaving is proposed. The approach was implemented on top of classical constraint propagation algorithms and tested on resource constrained project scheduling problems (RCPSP). We were able to close more than 15% of the previously open problems of the PSPLIB [Kolisch and Sprecher, 1996] and improve more than 31% of the best known lower bounds on those heavily studied problems. Other new results on open-shop and cumulative job-shop scheduling are reported.

Lecture Notes in Computer Science, 2003

The problem we tackle is progressive scheduling with temporal and resource uncertainty. Operation... more The problem we tackle is progressive scheduling with temporal and resource uncertainty. Operation durations are imprecise and alternative resources may break down. Operation end times and resource breakdowns are observed during execution. In this paper, we assume we have a representation of uncertainty in the form of probability distributions which are used in the simulation of schedule execution. We generate the schedule piece by piece during execution and use simulation to monitor the execution of the partial schedule. This paper describes the basis on which the decision to select and schedule a new subset of operations is made.

European Journal of Operational Research, 2003

A destructive lower bound for the multi-mode resource-constrained project scheduling problem with... more A destructive lower bound for the multi-mode resource-constrained project scheduling problem with minimal and maximal time-lags is presented. Given are n activities which may be processed in different modes without preemptions. During processing certain amounts of renewable and non-renewable resources are needed where the available capacity of each resource type is limited. Furthermore, minimal and maximal time-lags between the activities

Artificial Intelligence, 2003

This paper summarizes the main existing approaches to propagate resource constraints in Constrain... more This paper summarizes the main existing approaches to propagate resource constraints in Constraint-Based scheduling and identifies some of their limitations for using them in an integrated planning and scheduling framework. We then describe two new algorithms to propagate resource constraints on discrete resources and reservoirs. Unlike most of the classical work in scheduling, our algorithms focus on the precedence relations between activities rather than on their absolute position in time. They are efficient even when the set of activities is not completely defined and when the time window of activities is large. These features explain why our algorithms are particularly suited for integrated planning and scheduling approaches. All our algorithms are illustrated with examples. Encouraging preliminary results are reported on pure scheduling problems as well as some possible extensions of our framework.

This paper presents a Large Neighborhood Search (LNS) approach based on constraint programming to... more This paper presents a Large Neighborhood Search (LNS) approach based on constraint programming to solve cumulative scheduling problems. It extends earlier work on constraint-based randomized LNS for disjunctive scheduling as reported in (Nuijten & Le Pape 1998). A breakthrough development in generalizing that approach toward cumulative scheduling lies in the presented way of calculating a partial-order schedule from a fixed start time schedule. The approach is applied and tested on the Cumulative Job Shop Scheduling Problem (CJSSP). An empirical performance analysis is performed using a well-known set of benchmark instances. The described approach obtains the best known performance reported to date on the CJSSP. It not only finds better solutions than ever reported before for 33 out of 36 open instances, it also proves to be very robust on the complete set of test instances. Furthermore, among these 36 open instances, one is now closed. As the approach is generic, it can be applied to other types of scheduling problems, for example problems including resource types like reservoirs and state resources, and objectives like earliness/tardiness costs and resource allocation costs.

Providing robust scheduling algorithms that can solve a large variety of scheduling problems with... more Providing robust scheduling algorithms that can solve a large variety of scheduling problems with good performance is one of the biggest challenge of practical schedulers today. In this paper we present a robust scheduling algorithm based on Self-Adapting Large Neighborhood Search and apply it to a large panel of single-mode scheduling problems. The approach combines Large Neighborhood Search with a portfolio of neighborhoods and completion strategies together with Machine Learning techniques to converge on the most efficient neighborhoods and completion strategies for the problem being solved. The algorithm is evaluated on a set of 21 scheduling benchmarks, most of which are well established in the scheduling community. Despite the generality of the approach, for 17 benchmarks out of 21, its mean relative distance to state-of-the-art problem specific algorithms is less than 4%. It even outperforms state-of-the-art problem-specific algorithms on 7 benchmarks clearly showing that our...

This paper describes a simple complete search for cumulative scheduling based on the detection an... more This paper describes a simple complete search for cumulative scheduling based on the detection and resolution of minimal critical sets (MCS). The heuristic for selecting MCSs relies on an estimation of the related reduction of the search space. An extension of the search procedure using self-adapting shaving is proposed. The approach was implemented on top of classical constraint propagation algorithms and tested on resource constrained project scheduling problems (RCPSP). We were able to close more than 15% of the previously open problems of the PSPLIB [Kolisch and Sprecher, 1996] and improve more than 31% of the best known lower bounds on those heavily studied problems. Other new results on open-shop and cumulative job-shop scheduling are reported.

neighborhoods and completion strategies for the problem being solved. The algorithm is evaluated ... more neighborhoods and completion strategies for the problem being solved. The algorithm is evaluated on a set of 21 scheduling benchmarks, most of which are well established in the scheduling community. Despite the generality of the approach, for 17 benchmarks out of 21, its mean relative distance to state-of-the-art problem specific algorithms is less than 4%. It even outperforms state-of-the-art problem-specific algorithms on 7 benchmarks clearly showing that our algorithm oers a valuable compromise between robustness

You need to assign these 210 people to 12 groups of 16, 17, 18, or 19 people in every group. Of c... more You need to assign these 210 people to 12 groups of 16, 17, 18, or 19 people in every group. Of course, a person may be assigned to only one group. The objective is to maximize the diversity of people in the groups. Diversity means that it is preferable to assign people with the same characteristics to different groups. In this note we use a definition of diversity that is using a simple rule that measures the penalty when two people with similar characteristics are assigned to the same group and then minimize the total penalty. More precisely, the penalty for having two people with similar characteristics in the same group is equal to the number of characteristics they have in common. For instance if BIW and JRT are in the same group, it will incur a penalty of 2 because they have two characteristics in common: their department (NNE) and their gender (M). Thus the problem is to assign the people to the different groups while minimizing the total penalty.

HAL (Le Centre pour la Communication Scientifique Directe), 1997

International audienc

Proceedings of the International Conference on Automated Planning and Scheduling, Jun 5, 2017

Lecture Notes in Computer Science, 2018

HAL (Le Centre pour la Communication Scientifique Directe), Sep 8, 2014

Computers & Operations Research, Nov 1, 2020

Lecture Notes in Computer Science, 1999

Lecture Notes in Computer Science, 2018

International Joint Conference on Artificial Intelligence, 2005

Lecture Notes in Computer Science, 2003

European Journal of Operational Research, 2003

Artificial Intelligence, 2003

This paper describes a simple complete search for cumulative scheduling based on the detection an... more This paper describes a simple complete search for cumulative scheduling based on the detection and resolution of minimal critical sets (MCS). The heuristic for selecting MCSs relies on an estimation of the related reduction of the search space. An extension of the search procedure using self-adapting shaving is proposed. The approach was implemented on top of classical constraint propagation algorithms and tested on resource constrained project scheduling problems (RCPSP). We were able to close more than 15% of the previously open problems of the PSPLIB [Kolisch and Sprecher, 1996] and improve more than 31% of the best known lower bounds on those heavily studied problems. Other new results on open-shop and cumulative job-shop scheduling are reported.