Generating hard instances for robust combinatorial optimization

While research in robust optimization has attracted considerable interest over the last decades, its algorithmic development has been hindered by several factors. One of them is a missing set of benchmark instances that make algorithm performance better comparable, and makes reproducing instances unnecessary. Such a benchmark set should contain hard instances in particular, but so far, the standard approach to produce instances has been to sample values randomly from a uniform distribution.In this paper we introduce a new method to produce hard instances for min-max combinatorial optimization problems, which is based on an optimization model itself. Our approach does not make any assumptions on the problem structure and can thus be applied to any combinatorial problem. Using the Selection and Traveling Salesman problems as examples, we show that it is possible to produce instances which are up to 500 times harder to solve for a mixed-integer programming solver than the current state-of-the-art instances.     

Experimental analysis of heuristics for the bottleneck traveling salesman problem

In this paper we develop efficient heuristic algorithms to solve the bottleneck traveling salesman problem (BTSP). Results of extensive computational experiments are reported. Our heuristics produced optimal solutions for all the test problems considered from TSPLIB, JM-instances, National TSP instances, and VLSI TSP instances in very reasonable running time. We also conducted experiments with specially constructed ‘hard’ instances of the BTSP that produced optimal solutions for all but seven problems. Some fast construction heuristics are also discussed. Our algorithms could easily be modified to solve related problems such as the maximum scatter TSP and testing hamiltonicity of a graph.     

Ellipsoidal Approach to Box-Constrained Quadratic Problems

We present a new heuristic for the global solution of box constrained quadratic problems, based on the classical results which hold for the minimization of quadratic problems with ellipsoidal constraints. The approach is tested on several problems randomly generated and on graph instances from the DIMACS challenge, medium size instances of the Maximum Clique Problem. The numerical results seem to suggest some effectiveness of the proposed approach.     

RobinX: A three-field classification and unified data format for round-robin sports timetabling

Sports timetabling problems are combinatorial optimization problems which consist of creating a timetable that defines against whom, when, and where teams play games. In the literature, sports timetabling problems have been reported featuring a wide variety of constraints and objectives. This variety makes it challenging to identify the relevant set of papers for a given sports timetabling problem. Moreover, the lack of a generally accepted data format makes that problem instances and their solutions are rarely shared. Consequently, it is hard to assess algorithmic performance since solution methods are often tested on just one or two specific instances. To mitigate these issues, this paper presents RobinX, a three-field notation to describe a sports timetabling problem by means of the tournament format, the constraints in use, and the objective. We use this notation to classify sports timetabling problems presented in the operations research literature during the last five decades. Moreover, RobinX contains xml-based file templates to store problem instances and their solutions and presents an online platform that offers three useful tools. First, a query tool assists users to select the relevant set of papers for a given timetabling problem. Second, the online platform provides access to an xml data repository that contains real-life problem instances from different countries and sports. Finally, the website enables users to interact with a free and open-source C++-library to read and write xml files and to validate and evaluate encoded instances and solutions.     

A reduced variable neighborhood search algorithm for uncapacitated multilevel lot-sizing problems

Multilevel lot-sizing (MLLS) problems, which involve complicated product structures with interdependence among the items, play an important role in the material requirement planning (MRP) system of modern manufacturing/assembling lines. In this paper, we present a reduced variable neighborhood search (RVNS) algorithm and several implemental techniques for solving uncapacitated MLLS problems. Computational experiments are carried out on three classes of benchmark instances under different scales (small, medium, and large). Compared with the existing literature, RVNS shows good performance and robustness on a total of 176 tested instances. For the 96 small-sized instances, the RVNS algorithm can find 100% of the optimal solutions in less computational time; for the 40 medium-sized and the 40 large-sized instances, the RVNS algorithm is competitive against other methods, enjoying good effectiveness as well as high computational efficiency. In the calculations, RVNS updated 7 (17.5%) best known solutions for the medium-sized instances and 16 (40%) best known solutions for the large-sized instances.     

A general model for cyclic machine scheduling problems

A general framework for modeling and solving cyclic scheduling problems is presented. The objective is to minimize the cycle time. The model covers different cyclic versions of the job-shop problem found in the literature, robotic cell problems, the single hoist scheduling problem and tool transportation between the machines.It is shown that all these problems can be formulated as mixed integer linear programs which have a common structure. Small instances are solved with CPLEX. For larger instances tabu search procedures have been developed. The main ideas of these methods are indicated.     

Computational experience with an interior point algorithm on the satisfiability problem

We apply the zero-one integer programming algorithm described in Karmarkar [12] and Karmarkar, Resende and Ramakrishnan [13] to solve randomly generated instances of the satisfiability problem (SAT). The interior point algorithm is briefly reviewed and shown to be easily adapted to solve large instances of SAT. Hundreds of instances of SAT (having from 100 to 1000 variables and 100 to 32,000 clauses) are randomly generated and solved. For comparison, we attempt to solve the problems via linear programming relaxation with MINOS.     

Optimal berth allocation and time-invariant quay crane assignment in container terminals

Due to the dramatic increase in the world’s container traffic, the efficient management of operations in seaport container terminals has become a crucial issue. In this work, we focus on the integrated planning of the following problems faced at container terminals: berth allocation, quay crane assignment (number), and quay crane assignment (specific). First, we formulate a new binary integer linear program for the integrated solution of the berth allocation and quay crane assignment (number) problems called BACAP. Then we extend it by incorporating the quay crane assignment (specific) problem as well, which is named BACASP. Computational experiments performed on problem instances of various sizes indicate that the model for BACAP is very efficient and even large instances up to 60 vessels can be solved to optimality. Unfortunately, this is not the case for BACASP. Therefore, to be able to solve large instances, we present a necessary and sufficient condition for generating an optimal solution of BACASP from an optimal solution of BACAP using a post-processing algorithm. In case this condition is not satisfied, we make use of a cutting plane algorithm which solves BACAP repeatedly by adding cuts generated from the optimal solutions until the aforementioned condition holds. This method proves to be viable and enables us to solve large BACASP instances as well. To the best of our knowledge, these are the largest instances that can be solved to optimality for this difficult problem, which makes our work applicable to realistic problems.     

A biased random-key genetic algorithm for the Steiner triple covering problem

We present a biased random-key genetic algorithm (BRKGA) for finding small covers of computationally difficult set covering problems that arise in computing the 1-width of incidence matrices of Steiner triple systems. Using a parallel implementation of the BRKGA, we compute improved covers for the two largest instances in a standard set of test problems used to evaluate solution procedures for this problem. The new covers for instances A ₄₀₅ and A ₇₂₉ have sizes 335 and 617, respectively. On all other smaller instances our algorithm consistently produces covers of optimal size.     

A cutting-plane approach for large-scale capacitated multi-period facility location using a specialized interior-point method

We propose a cutting-plane approach (namely, Benders decomposition) for a class of capacitated multi-period facility location problems. The novelty of this approach lies on the use of a specialized interior-point method for solving the Benders subproblems. The primal block-angular structure of the resulting linear optimization problems is exploited by the interior-point method, allowing the (either exact or inexact) efficient solution of large instances. The consequences of different modeling conditions and problem specifications on the computational performance are also investigated both theoretically and empirically, providing a deeper understanding of the significant factors influencing the overall efficiency of the cutting-plane method. The methodology proposed allowed the solution of instances of up to 200 potential locations, one million customers and three periods, resulting in mixed integer linear optimization problems of up to 600 binary and 600 millions of continuous variables. Those problems were solved by the specialized approach in less than one hour and a half, outperforming other state-of-the-art methods, which exhausted the (144 GB of) available memory in the largest instances.     

A probabilistic heuristic for a computationally difficult set covering problem

An efficient probabilistic set covering heuristic is presented. The heuristic is evaluated on empirically difficult to solve set covering problems that arise from Steiner triple systems. The optimal solution to only a few of these instances is known. The heuristic provides these solutions as well as the best known solutions to all other instances attempted.     

Isomorphic scheduling problems

We consider general properties of isomorphic scheduling problems that constitute a new class of pairs of mutually related scheduling problems. Any such a pair is composed of a scheduling problem with fixed job processing times and its time-dependent counterpart with processing times that are proportional-linear functions of the job starting times. In order to introduce the class formally, first we formulate a generic scheduling problem with fixed job processing times and define isomorphic problems by a one-to-one transformation of instances of the generic problem into instances of time-dependent scheduling problems with proportional-linear job processing times. Next, we prove basic properties of isomorphic scheduling problems and show how to convert polynomial algorithms for scheduling problems with fixed job processing times into polynomial algorithms for proportional-linear counterparts of the original problems. Finally, we show how are related approximation algorithms for isomorphic problems. Applying the results, we establish new worst-case results for time-dependent parallel-machine scheduling problems and prove that many single- and dedicated-machine time-dependent scheduling problems with proportional-linear job processing times are polynomially solvable.     

A comparison of two methods for solving 0–1 integer programs using a general purpose simulated annealing algorithm

0–1 problems are often difficult to solve. Although special purpose algorithms (exact as well as heuristic) exist for solving particular problem classes or problem instances, there are few general purpose algorithms for solving practical-sized instances of 0–1 problems. This paper deals with a general purpose heuristic algorithm for 0–1 problems. In this paper, we compare two methods based on simulated annealing for solving general 0–1 integer programming problems. The two methods differe in the scheme used for neighbourhood transitions in the simulated annealing framework. We compare the performance of the two methods on the set partitioning problem.     

Some tractable instances of interval data minmax regret problems

In this paper, we provide polynomial and pseudopolynomial algorithms for classes of particular instances of interval data minmax regret graph problems. These classes are defined using a parameter that measures the distance from well-known solvable instances. Tractable cases occur when the parameter is bounded by a constant.     

Semi-Lagrangian relaxation applied to?the?uncapacitated facility location problem

We show how the performance of general purpose Mixed Integer Programming (MIP) solvers, can be enhanced by using the Semi-Lagrangian Relaxation (SLR) method. To illustrate this procedure we perform computational experiments on large-scale instances of the Uncapacitated Facility Location (UFL) problems with unknown optimal values. CPLEX solves 3 out of the 36 instances. By combining CPLEX with SLR, we manage to solve 18 out of the 36 instances and improve the best known lower bound for the other instances. The key point has been that, on average, the SLR approach, has reduced by more than 90% the total number of relevant UFL variables.     

A hybrid imperialist competitive algorithm for minimizing makespan in a multi-processor open shop

In this paper a new mixed-integer linear programming (MILP) model is proposed for the multi-processor open shop scheduling (MPOS) problems to minimize the makespan with considering independent setup time and sequence dependent removal time. A hybrid imperialist competitive algorithm (ICA) with genetic algorithm (GA) is presented to solve this problem. The parameters of the proposed algorithm are tuned by response surface methodology (RSM). The performance of the algorithm to solve small, medium and large sized instances of the problem is evaluated by introducing two performance metrics. The quality of obtained solutions is compared with that of the optimal solutions for small sized instances and with the lower bounds for medium sized instances. Also some computational results are presented for large sized instances.     

A cross entropy based algorithm for reliability problems

The Cross Entropy method has recently been applied to combinatorial optimization problems with promising results. This paper proposes a Cross Entropy based algorithm for reliability optimization of complex systems, where one wants to maximize the reliability of a system through optimal allocation of redundant components while respecting a set of budget constraints. We illustrate the effectiveness of the proposed algorithm on two classes of problems, software system reliability optimization and complex network reliability optimization, by testing it on instances from the literature as well as on randomly generated large scale instances. Furthermore, we show how a Cross Entropy-based algorithm can be fine-tuned by using a training scheme based upon the Response Surface Methodology. Computational results show the effectiveness as well as the robustness of the algorithm on different classes of problems.     

A branch-and-bound-based local search method for the flow shop problem

It is well-known that exact branch and bound methods can only solve small or moderately sized ????-hard combinatorial optimization problems. In this paper, we address the issue of embedding an approximate branch and bound algorithm into a local search framework. The resulting heuristic has been applied to the problem of finding a minimum makespan in the permutation flow shop problem. Computational experiments carried out on a large set of benchmark problems show that the proposed method consistently yields optimal or near-optimal solutions for instances with up to 200 jobs and 10 machines. In particular, for 19 instances, the heuristic produces solutions that outperform the best known ones.     

An efficient grouping genetic algorithm for U-shaped assembly line balancing problems with maximizing production rate

U-type assembly line is one of the important tools that may increase companies’ production efficiency. In this study, two different modeling approaches proposed for the assembly line balancing problems have been used in modeling type-II U-line balancing problems, and the performances of these models have been compared with each other. It has been shown that using mathematical formulations to solve medium and large size problem instances is impractical since the problem is NP-hard. Therefore, a grouping genetic and simulated annealing algorithms have been developed, and a particle swarm optimization algorithm is adapted to compare with the proposed methods. A special crossover operator that always obtains feasible offspring has been suggested for the proposed grouping genetic algorithm. Furthermore, a local search procedure based on problem-specific knowledge was applied to increase the intensification of the algorithm. A set of well-known benchmark instances was solved to evaluate the effectiveness of the proposed and existing methods. Results showed that while the mathematical formulations can only be used to solve small size instances, metaheuristics can obtain high quality solutions for all size problem instances within acceptable CPU times. Moreover, grouping genetic algorithm has been found to be superior to the other methods according to the number of optimal solutions, or deviations from the lower bound values.     

On multi-item economic lot-sizing with remanufacturing and uncapacitated production

In this study, we consider the multi-item economic lot-sizing problem with remanufacturing and uncapacitated production. By observing that the problem comprises several independent single-item problems, we show how very high quality feasible solutions and bounds can be obtained by solving each item separately using an effective recently proposed approach. Computational experiments demonstrate that our approach improves the best known feasible solutions and lower bounds for all the available instances. In addition, we show that 86 instances can be solved to optimality and the remaining open gap is below 0.5% for almost all the unsolved instances.