Accelerating column generation for aircraft scheduling using constraint propagation

We discuss how constraint programming can improve the performance of a column generation solution process for the NP-hard Tail Assignment problem in aircraft scheduling. Combining a constraint model of a relaxed Tail Assignment problem with column generation, we achieve substantially improved performance. A generalized preprocessing technique based on constraint propagation is presented that can dramatically reduce the size of the flight network. We also present a heuristic preprocessing method based on the costs of connections, and show how constraint propagation can be used to improve fixing heuristics. Proof of concept is provided using real world Tail Assignment instances.     

Solving an integrated employee timetabling and job-shop scheduling problem via hybrid branch-and-bound

We propose exact hybrid methods based on integer linear programming (ILP) and constraint programming (CP) for an integrated employee timetabling and job-shop scheduling problem. Each method we investigate uses a CP formulation associated with an LP relaxation. Under a CP framework, the LP relaxation is integrated into a global constraint using in addition reduced cost-based filtering techniques. We propose two CP formulations of the problem yielding two different LP relaxations. The first formulation is based on a direct representation of the problem. The second formulation is based on a decomposition in intervals of the possible operation starting times. We show the theoretical interest of the decomposition-based representation compared to the direct representation through the analysis of dominant schedules. Computational experiments on a set of randomly generated instances confirm the superiority of the decomposition-based representation. In both cases, the hybrid methods outperform pure CP for employee cost minimization while it is not the case for makespan minimization. The experiments also investigate the interest of the proposed integrated method compared to a sequential approach and show its potential for multiobjective optimization.     

Solving subgraph isomorphism problems with constraint programming

The subgraph isomorphism problem consists in deciding if there exists a copy of a pattern graph in a target graph. We introduce in this paper a global constraint and an associated filtering algorithm to solve this problem within the context of constraint programming. The main idea of the filtering algorithm is to label every node with respect to its relationships with other nodes of the graph, and to define a partial order on these labels in order to express compatibility of labels for subgraph isomorphism. This partial order over labels is used to filter domains. Labelings can also be strengthened by adding information from the labels of neighbors. Such a strengthening can be applied iteratively until a fixpoint is reached. Practical experiments illustrate that our new filtering approach is more effective on difficult instances of scale free graphs than state-of-the-art algorithms and other constraint programming approaches.     

New Hybrid Optimization Algorithms for Machine Scheduling Problems

Dynamic programming, branch-and-bound, and constraint programming are the standard solution principles for finding optimal solutions to machine scheduling problems. We propose a new hybrid optimization framework that integrates all three methodologies. The hybrid framework leads to powerful solution procedures. We demonstrate our approach through the optimal solution of the single-machine total weighted completion time scheduling problem subject to release dates, which is known to be strongly NP-hard. Extensive computational experiments indicate that new hybrid algorithms use orders of magnitude less storage than dynamic programming, and yet can still reap the full benefit of the dynamic programming property inherent to the problem. We are able to solve to optimality all 1900 instances with up to 200 jobs. This more than doubles the size of problems that can be solved optimally by the previous best algorithm running on the latest computing hardware.     

Scheduling parallel-machine batch operations to maximize on-time delivery performance

In this paper we study the problem of minimizing total weighted tardiness, a proxy for maximizing on-time delivery performance, on parallel nonidentical batch processing machines. We first formulate the (primal) problem as a nonlinear integer programming model. We then show that the primal problem can be solved exactly by solving a corresponding dual problem with a nonlinear relaxation. Since both the primal and the dual problems are NP-hard, we use genetic algorithms, based on random keys and multiple choice encodings, to heuristically solve them. We find that the genetic algorithms consistently outperform a standard mathematical programming package in terms of solution quality and computation time. We also compare the smaller problem instances to a breadth-first tree search algorithm that gives evidence of the quality of the solutions.     

An exact algorithm for the single-machine total weighted tardiness problem with sequence-dependent setup times

This study proposes an exact algorithm for the single-machine total weighted tardiness problem with sequence-dependent setup times. The algorithm is an extension of the authors' previous algorithm for the single-machine scheduling problem without setup times, which is based on the SSDP (Successive Sublimation Dynamic Programming) method. In the first stage of the algorithm, the conjugate subgradient algorithm or the column generation algorithm is applied to a Lagrangian relaxation of the original problem to adjust multipliers. Then, in the second stage, constraints are successively added to the relaxation until the gap between lower and upper bounds becomes zero. The relaxation is solved by dynamic programming and unnecessary dynamic programming states are eliminated to suppress the increase of computation time and memory space. In this study a branching scheme is integrated into the algorithm to manage to solve hard instances. The proposed algorithm is applied to benchmark instances in the literature and almost all of them are optimally solved.     

The patient bed assignment problem solved by autonomous bat algorithm

The patient bed assignment problem consists of managing, in the best possible way, a set of beds with particular features and assigning them to a set of patients with special requirements. This assignment problem can be seen an optimization problem, of which the intended aims are usually to minimize the number of internal movements within a unit and to maximize bed usage according to the levels of criticality of the patients, among others. The usual approaches for solving this problem follow a traditional model based on the constraint programming paradigm, mainly using hard constraints. However, in real-life problems, constraints that should ideally be satisfied are often violated. In this paper, we present a new model for the patient bed assignment problem based on the minimum sum of unsatisfied constraints. This technique enables the consideration of soft constraints in the potential solutions that exhibit the best performance. The aim is to find the assignment that minimizes a weighted sum of the unsatisfied constraints. To this end, we use an autonomous binary version of the bat algorithm, which is an optimization technique inspired by the bio-sonar behaviour of microbats, to find the best set of potential solutions without requiring any expert user knowledge to achieve an efficient solution process. To validate our proposal, we use our model to solve problem instances based on data from several hospitals, and we perform a detailed comparative statistical analysis with a traditional constraint programming solver and several well-known optimization algorithms, including the classic bat algorithm. Promising results show that our approach is capable of efficiently solving 30 instances with decreased solution times.     

A hybrid genetic and linear programming algorithm for two-agent order acceptance and scheduling problem

In this paper, the simultaneous order acceptance and scheduling problem is developed by considering the variety of customers’ requests. To that end, two agents with different scheduling criteria including the total weighted lateness for the first and the weighted number of tardy orders for the second agent are considered. The objective is to maximize the sum of the total profit of the first and the total revenue of the second agents’ orders when the weighted number of tardy orders of the second agent is bounded by an upper bound value. In this study, it is shown that this problem is NP-hard in the strong sense, and then to optimally solve it, an integer linear programming model is proposed based on the properties of optimal solution. This model is capable of solving problem instances up to 60 orders in size. Also, the LP-relaxation of this model was used to propose a hybrid meta-heuristic algorithm which was developed by employing genetic algorithm and linear programming. Computational results reveal that the proposed meta-heuristic can achieve near optimal solutions so efficiently that for the instances up to 60 orders in size, the average deviation of the model from the optimal solution is lower than 0.2% and for the instances up to 150 orders in size, the average deviation from the problem upper bound is lower than 1.5%.     

Improved Approximations for Guarding 1.5-Dimensional Terrains

We present a 4-approximation algorithm for the problem of placing the fewest guards on a 1.5D terrain so that every point of the terrain is seen by at least one guard. This improves on the previous best approximation factor of 5 (see King in Proceedings of the 13th Latin American Symposium on Theoretical Informatics, pp. 629–640, 2006). Unlike most of the previous techniques, our method is based on rounding the linear programming relaxation of the corresponding covering problem. Besides the simplicity of the analysis, which mainly relies on decomposing the constraint matrix of the LP into totally balanced matrices, our algorithm, unlike previous work, generalizes to the weighted and partial versions of the basic problem.     

Efficient filtering for the Resource-Cost AllDifferent constraint

This paper studies a family of optimization problems where a set of items, each requiring a possibly different amount of resource, must be assigned to different slots for which the price of the resource can vary. The objective is then to assign items such that the overall resource cost is minimized. Such problems arise commonly in domains such as production scheduling in the presence of fluctuating renewable energy costs or variants of the Travelling Salesman Problem. In Constraint Programming, this can be naturally modeled in two ways: (a) with a sum of element constraints; (b) with a MinimumAssignment constraint. Unfortunately the sum of element constraints obtains a weak filtering and the MinimumAssignment constraint does not scale well on large instances. This work proposes a third approach by introducing the ResourceCostAllDifferent constraint and an associated incremental and scalable filtering algorithm, running in \(\mathcal {O}(n \cdot m)\), where n is the number of unbound variables and m is the maximum domain size of unbound variables. Its goal is to compute the total cost in a scalable manner by dealing with the fact that all assignments must be different. We first evaluate the efficiency of the new filtering on a real industrial problem and then on the Product Matrix Travelling Salesman Problem, a special case of the Asymmetric Travelling Salesman Problem. The study shows experimentally that our approach generally outperforms the decomposition and the MinimumAssignment ones for the problems we considered.     

Lower bounds and exact algorithms for the quadratic minimum spanning tree problem

We address the quadratic minimum spanning tree problem (QMSTP), the problem of finding a spanning tree of a connected and undirected graph such that a quadratic cost function is minimized. We first propose an integer programming formulation based on the reformulation–linearization technique (RLT). We then use the idea of partitioning spanning trees into forests of a given fixed size and obtain a QMSTP reformulation that generalizes the RLT model. The reformulation is such that the larger the size of the forests, the stronger lower bounds provided. Thus, a hierarchy of formulations is obtained. At the lowest hierarchy level, one has precisely the RLT formulation, which is already stronger than previous formulations in the literature. The highest hierarchy level provides the convex hull of integer feasible solutions for the problem. The formulations introduced here are not compact, so the direct evaluation of their linear programming relaxation bounds is not practical. To overcome that, we introduce two lower bounding procedures based on Lagrangian relaxation. These procedures are embedded into two parallel branch-and-bound algorithms. As a result of our study, several instances in the literature were solved to optimality for the first time.     

Sequencing a single machine with due dates and deadlines: an ILP-based approach to solve very large instances

We consider the problem of minimizing the weighted number of tardy jobs on a single machine where each job is also subject to a deadline that cannot be violated. We propose an exact method based on a compact integer linear programming formulation of the problem and an effective reduction procedure that allows to solve to optimality instances with up to 30,000 jobs in size, and up to 50,000 jobs in size for the special deadline-free case.     

Models for a traveling purchaser problem with additional side-constraints

The traveling purchaser problem (TPP) is the problem of determining a tour of a purchaser that needs to buy several items in different shops such that the total amount of travel and purchase costs is minimized. Motivated by an application in machine scheduling, we study a variant of the problem with additional constraints, namely, a limit on the maximum number of markets to be visited, a limit on the number of items bought per market and where only one copy per item needs to be bought. We present an integer linear programming (ILP) model which is adequate for obtaining optimal integer solutions for instances with up to 100 markets. We also present and test several variations of a Lagrangian relaxation combined with a subgradient optimization procedure. The relaxed problem can be solved by dynamic programming and can also be viewed as resulting from applying a state space relaxation technique to a dynamic programming formulation. The Lagrangian based method is combined with a heuristic that attempts to transform relaxed solutions into feasible solutions. Computational results for instances with up to 300 markets show that with the exception of a few cases, the reported differences between best upper bound and lower bound values on the optimal solutions are reasonably small.     

A New Approach for Weighted Constraint Satisfaction

We consider the Weighted Constraint Satisfaction Problem which is an important problem in Artificial Intelligence. Given a set of variables, their domains and a set of constraints between variables, our goal is to obtain an assignment of the variables to domain values such that the weighted sum of satisfied constraints is maximized. In this paper, we present a new approach based on randomized rounding of semidefinite programming relaxation. Besides having provable worst-case bounds for domain sizes 2 and 3, our algorithm is simple and efficient in practice, and produces better solutions than some other polynomial-time algorithms such as greedy and randomized local search.     

Complexity of Clausal Constraints Over Chains

We investigate the complexity of the satisfiability problem of constraints over finite totally ordered domains. In our context, a clausal constraint is a disjunction of inequalities of the form x≥d and x≤d. We classify the complexity of constraints based on clausal patterns. A pattern abstracts away from variables and contains only information about the domain elements and the type of inequalities occurring in a constraint. Every finite set of patterns gives rise to a (clausal) constraint satisfaction problem in which all constraints in instances must have an allowed pattern. We prove that every such problem is either polynomially decidable or NP-complete, and give a polynomial-time algorithm for recognizing the tractable cases. Some of these tractable cases are new and have not been previously identified in the literature.     

Approximated consistency for the automatic recording constraint

We introduce the automatic recording constraint (ARC) that can be used to model and solve scheduling problems where tasks may not overlap in time and the tasks linearly exhaust some resource. Since achieving generalized arc-consistency for the ARC is NP-hard, we develop a filtering algorithm that achieves approximated consistency only. Numerical results show the benefits of the new constraint on three out of four different types of benchmark sets for the automatic recording problem. On these instances, run-times can be achieved that are orders of magnitude better than those of the best previous constraint programming approach.     

A global constraint for total weighted completion time for unary resources

We introduce a novel global constraint for the total weighted completion time of activities on a single unary capacity resource. For propagating the constraint, we propose an O(n ⁴) algorithm which makes use of the preemptive mean busy time relaxation of the scheduling problem. The solution to this problem is used to test if an activity can start at each start time in its domain in solutions that respect the upper bound on the cost of the schedule. Empirical results show that the proposed global constraint significantly improves the performance of constraint-based approaches to single-machine scheduling for minimizing the total weighted completion time. We then apply the constraint to the multi-machine job shop scheduling problem with total weighted completion time. Our experiments show an order of magnitude reduction in search effort over the standard weighted-sum constraint and demonstrate that the way in which the job weights are associated with activities is important for performance.     

A minisum location problem with regional demand considering farthest Euclidean distances

We consider a continuous multi-facility location-allocation problem that aims to minimize the sum of weighted farthest Euclidean distances between (closed convex) polygonal and/or circular demand regions, and facilities they are assigned to. We show that the single facility version of the problem has a straightforward second-order cone programming formulation and can therefore be efficiently solved to optimality. To solve large size instances, we adapt a multi-dimensional direct search descent algorithm to our problem which is not guaranteed to find the optimal solution. In a special case with circular and rectangular demand regions, this algorithm, if converges, finds the optimal solution. We also apply a simple subgradient method to the problem. Furthermore, we review the algorithms proposed for the problem in the literature and compare all these algorithms in terms of both solution quality and time. Finally, we consider the multi-facility version of the problem and model it as a mixed integer second-order cone programming problem. As this formulation is weak, we use the alternate location-allocation heuristic to solve large size instances.     

A solution approach based on Benders decomposition for the preventive maintenance scheduling problem of a stochastic large-scale energy system

This paper describes a Benders decomposition-based framework for solving the large scale energy management problem that was posed for the ROADEF 2010 challenge. The problem was taken from the power industry and entailed scheduling the outage dates for a set of nuclear power plants, which need to be regularly taken down for refueling and maintenance, in such a way that the expected cost of meeting the power demand in a number of potential scenarios is minimized. We show that the problem structure naturally lends itself to Benders decomposition; however, not all constraints can be included in the mixed integer programming model. We present a two phase approach that first uses Benders decomposition to solve the linear programming relaxation of a relaxed version of the problem. In the second phase, integer solutions are enumerated and a procedure is applied to make them satisfy constraints not included in the relaxed problem. To cope with the size of the formulations arising in our approach we describe efficient preprocessing techniques to reduce the problem size and show how aggregation can be applied to each of the subproblems. Computational results on the test instances show that the procedure competes well on small instances of the problem, but runs into difficulty on larger ones. Unlike heuristic approaches, however, this methodology can be used to provide lower bounds on solution quality.