Constraint Programming and Operations Research: Comments from an Operations Researcher

This paper relates the author's personal experience with constraint programming and gives a personal assessment of the relationships between constraint programming and operations research.     

Crew Assignment via Constraint Programming: Integrating Column Generation and Heuristic Tree Search

The Airline Crew Assignment Problem (ACA) consists of assigning lines of work to a set of crew members such that a set of activities is partitioned and the costs for that assignment are minimized. Especially for European airline companies, complex constraints defining the feasibility of a line of work have to be respected. We developed two different algorithms to tackle the large scale optimization problem of Airline Crew Assignment. The first is an application of the Constraint Programming (CP) based Column Generation Framework. The second approach performs a CP based heuristic tree search. We present how both algorithms can be coupled to overcome their inherent weaknesses by integrating methods from Constraint Programming and Operations Research. Numerical results show the superiority of the hybrid algorithm in comparison to CP based tree search and column generation alone.     

Mathematical Programming Techniques in Constraint Programming: A Short Overview

In recent years, the integration of techniques from Artificial Intelligence and Operations Research has shown to improve the solutions of complex and large scale combinatorial optimization problems, in terms of efficiency, scalability and optimality. In this context, Constraint Programming is an emerging discipline situated at the confluence of the two fields that has been recognized as a suitable environment for achieving such an integration. This paper briefly presents the integration directions explored in the literature, and provides some pointers to relevant work in these directions.     

运筹学中若干离散规划问题的相对差分搜索解法

运筹学中有很多离散规划问题。其中的线性规划通常用分枝定界法或割平面法，还有图上作业法求解。不论哪种方法工作量都不小，而且效率低；至于非线性规划大都是用动态规划法求解，也很麻烦、耗时。对于大规模问题，不论线性或非线性离散规划，现有解法都受到问题规模的限制；还有资源分配和背包问题至今没有见到解决方法。本文就是为了解决这些问题，提出了相对差分搜索算法。通过5个算例和其它文献中的一些算例计算验证了本法简单、快速、有效和精确，尤其不受问题规模的限制是其最大的优点。     

Encoding Binary Arithmetic Operations in Integer Programming Formulations

This paper presents the encoding of binary arithmetic operations within Integer Programming (IP) formulations, specifically the encoding of binary multiplication and addition/subtraction. This allows the direct manipulation of integer quantities represented as binary strings of arbitrary size. Many articles published in the past within the Chemical Engineering community have used this representation of integer quantities within Mixed-Integer formulations for Process Optimization and Design. Applications such as these can benefit from the formulations derived in this work. As a demonstrative application we consider the simple number factorization problem, according to which given an odd number C factors A and B are to be found such that C equals their product. If any such factors are found the number is factorable, else it is proven to be prime. An IP formulation is derived involving upper and lower bounding logical constraints to encode for the value of the binary string digits. The formulation involves \({\cal O}(\log C)\) binary variables, \({\cal O}((\log C)^{2})\) continuous variables, and \({\cal O}((\log C)^{2})\) constraints to describe the problem. Computational results demonstrate the validity of this approach, highlighting also the fact that such formulations are not very tight thus resulting in large numbers of iterations of the Branch and Bound algorithm used. It is also observed that the formulations become significantly tighter if logical upper bounding constraints forcing continuous variables involved to be zero are included.     

较多约束规划及其最优性条件

本文研究了文[1]中提出的较多约束规划的基本问题，在给出问题的较多约束集结构表示的基础上，证明了这类问题的较多约束最优解要满足的Fritz John条件和Kuhn-Tucker条件。     

A Functorial Framework for Constraint Normal Logic Programming

The semantic constructions and results for definite programs do not extend when dealing with negation. The main problem is related to a well-known problem in the area of algebraic specification: if we fix a constraint domain as a given model, its free extension by means of a set of Horn clauses defining a set of new predicates is semicomputable. However, if the language of the extension is richer than Horn clauses its free extension (if it exists) is not necessarily semicomputable. In this paper we present a framework that allows us to deal with these problems in a novel way. This framework is based on two main ideas: a reformulation of the notion of constraint domain and a functorial presentation of our semantics. In particular, the semantics of a logic program P is defined in terms of three functors: that apply to constraint domains and provide the operational, the least fixpoint and the logical semantics of P, respectively. To be more concrete, the idea is that the application of to a specific constraint solver provides the operational semantics of P that uses this solver; the application of to a specific domain provides the least fixpoint of P over this domain; and, the application of to a theory of constraints provides the logic theory associated to P. In this context, we prove that these three functors are in some sense equivalent.      

Solving VRPTWs with Constraint Programming Based Column Generation

Constraint programming based column generation is a hybrid optimization framework recently proposed (Junker et al., 1999) that uses constraint programming to solve column generation subproblems. In the past, this framework has been used to solve scheduling problems where the associated graph is naturally acyclic and has done so very efficiently. This paper attempts to solve problems whose graph is cyclic by nature, such as routing problems, by solving the elementary shortest path problem with constraint programming. We also introduce new redundant constraints which can be useful in the general framework. The experimental results are comparable to those of the similar method in the literature (Desrochers, Desrosiers, and Solomon, 1992) but the proposed method yields a much more flexible approach.     

Constraint Programming Based Column Generation for Crew Assignment

Airline crew assignment problems are large-scale optimization problems which can be adequately solved by column generation. The subproblem is typically a so-called constrained shortest path problem and solved by dynamic programming. However, complex airline regulations arising frequently in European airlines cannot be expressed entirely in this framework and limit the use of pure column generation. In this paper, we formulate the subproblem as a constraint satisfaction problem, thus gaining high expressiveness. Each airline regulation is encoded by one or several constraints. An additional constraint which encapsulates a shortest path algorithm for generating columns with negative reduced costs is introduced. This constraint reduces the search space of the subproblem significantly. Resulting domain reductions are propagated to the other constraints which additionally reduces the search space. Numerical results based on data of a large European airline are presented and demonstrate the potential of our approach.     

A Constraint Programming Framework for Local Search Methods

We propose in this paper a novel integration of local search algorithms within a constraint programming framework for combinatorial optimization problems, in an attempt to gain both the efficiency of local search methods and the flexibility of constraint programming while maintaining a clear separation between the constraints of the problem and the actual search procedure. Each neighborhood exploration is performed by branch-and-bound search, whose potential pruning capabilities open the door to more elaborate local moves, which could lead to even better approximate results. Two illustrations of this framework are provided, including computational results for the traveling salesman problem with time windows. These results indicate that it is one order of magnitude faster than the customary constraint programming approach to local search and that it is competitive with a specialized local search algorithm.     

Combining Constraint Programming and Linear Programming on an Example of Bus Driver Scheduling

This paper provides details of a successful application where the Column Generation algorithm was used to combine Constraint Programming and Linear Programming. In the past, constraint programming and linear programming were considered to be two competing technologies that solved similar types of problems. Both these technologies had their strengths and weaknesses. This paper shows that the two technologies can be combined together to extract the strengths of both these technologies. Details of a real-world application to optimize bus driver duties is given here. This system was developed by ILOG for a major software house in Japan using ILOG-Solver and ILOG-CPLEX, constraint programming and linear programming C/C++ libraries.     

Mixed Duality without Constraint Qualification for Nonsmooth Fractional Programming

A class of nonsmooth multiobjective fractional programming is formulated. We establish the necessary and sufficient optimality conditions without the need of a constraint qualification. Then a mixed dual is introduced for a class of nonsmooth fractional programming problems, and various duality theorems are established without a constraint qualification.     

一类稀疏约束非线性规划的约束规格

研究一类带有闭凸集约束的稀疏约束非线性规划问题,这类问题在变量选择、模式识别、投资组合等领域具有广泛的应用.首先引进了限制性Slater约束规格的概念,证明了该约束规格强于限制性M-F约束规格,然后在此约束规格成立的条件下,分析了其局部最优解成立的充分和必要条件.最后,对约束集合的两种具体形式,指出限制性Slater约束规格必满足,并给出了一阶必要性条件的具体表达形式.     

Duality Without a Constraint Qualification for Minimax Fractional Programming

Using a parametric approach, we establish the necessary and sufficient conditions for generalized fractional programming without the need of a constraint qualification. Subsequently, these optimality criteria are utilized as a basis for constructing a parametric dual model and two other parameter-free dual models. Several duality theorems are established.     

Solving Vehicle Routing Problems Using Constraint Programming and Metaheuristics

Constraint Programming typically uses the technique of depth-first branch and bound as the method of solving optimization problems. Although this method can give the optimal solution, for large problems, the time needed to find the optimal can be prohibitive. This paper introduces a method for using local search techniques within a Constraint Programming framework, and applies this technique to vehicle routing problems. We introduce a Constraint Programming model for vehicle routing, and a system for integrating Constraint Programming and local search techniques. We then describe how the method can be accelerated by handling core constraints using fast local checks, while other more complex constraints are left to the constraint propagation system. We have coupled our local search method with a meta-heuristic to avoid the search being trapped in local minima. Several meta-heuristics are investigated ranging from a simple Tabu Search method to Guided Local Search. An empirical study over benchmark problems shows the relative merits of these techniques. Investigations indicate that the specific long-term memory technique used by Guided Local Search can be used as a diversification method for Tabu Search, resulting in significant benefit. Several new best solutions on the Solomon problems are found in relatively few iterations of our algorithm.     

Soft Constraint Logic Programming and Generalized Shortest Path Problems

In this paper we study the relationship between Constraint Programming (CP) and Shortest Path (SP) problems. In particular, we show that classical, multicriteria, partially ordered, and modality-based SP problems can be naturally modeled and solved within the Soft Constraint Logic Programming (SCLP) framework, where logic programming is coupled with soft constraints. In this way we provide this large class of SP problems with a high-level and declarative linguistic support whose semantics takes care of both finding the cost of the shortest path(s) and also of actually finding the path(s). On the other hand, some efficient algorithms for certain classes of SP problems can be exploited to provide some classes of SCLP programs with an efficient way to compute their semantics.     

Constraint Programming and Hybrid Formulations for Three Life Designs

Conway's game of Life provides an interesting testbed for exploring issues in formulation, symmetry, and optimization with constraint programming and hybrid constraint programming/integer programming methods. We consider three Life pattern-creation problems: finding maximum density still-Lifes, finding smallest immediate predecessor patterns, and finding period-2 oscillators. For the first two problems, integrating integer programming and constraint programming approaches provides a much better solution procedure than either individually. For the final problem, the constraint programming formulation provides the better approach.