Assessing a Metaheuristic for Large-Scale Commercial Districting

In this article, we present an empirical evaluation of a metaheuristic approach to a commercial districting problem. The problem consists of partitioning a given set of basic units into p districts in order to minimize a measure of territory dispersion. Additional constraints include territory connectivity and balancing with respect to several criteria. To obtain feasible solutions to this NP-hard problem, a reactive greedy randomized adaptive search metaheuristic procedure (GRASP) is used. Previous work addressed medium-scale instances. In this study, we report our computational experience when we addressed larger instances ressembling more closely the size of real-world instances. The empirical work includes full assessment of the algorithmic parameters and the local search phase, and a sensitivity analysis of the balance tolerance parameter in terms of solution quality and feasibility. The empirical evidence shows the effectiveness of the proposed approach and how this approach is significantly better than the method used by the industrial partner. The complexity of the planning constraints make the current practice method struggle to obtain feasible designs. Even for the larger cases, the proposed procedure successfuly solved instances with balance tolerance parameter values of as low as 3%, something impossible to achieve by the company’s current standards.     

GRASP with path relinking for commercial districting

The problem of grouping basic units into larger geographic territories subject to dispersion, connectivity, and balance requirements is addressed. The problem is motivated by a real-world application from the bottled beverage distribution industry. Typically, a dispersion function is minimized as compact territories are sought. Existing literature reveals that practically all the works on commercial districting use center-based dispersion functions. These center-based functions yield mixed-integer programming models with some nice properties; however, they have the disadvantage of being very costly to be properly evaluated when used within heuristic frameworks. This is due to the center updating operations frequently needed through the heuristic search. In this work, a more robust dispersion measure based on the diameter of the formed territories is studied. This allows a more efficient heuristic search computation. For solving this particular territory design problem, a greedy randomized adaptive search procedure (GRASP) that incorporates a novel construction procedure where territories are formed simultaneously in two main stages using different criteria is proposed. This also differs from previous literature where GRASP was used to build one territory at a time. The GRASP is further enhanced with two variants of forward-backward path relinking, namely static and dynamic. Path relinking is a sophisticated and very successful search mechanism. This idea is novel in any districting or territory design application to the best of our knowledge. The proposed algorithm and its components have been extensively evaluated over a wide set of data instances. Experimental results reveal that the construction mechanism produces feasible solutions of acceptable quality, which are improved by an effective local search procedure. In addition, empirical evidence indicate that the two path relinking strategies have a significant impact on solution quality when incorporated within the GRASP framework. The ideas and components of the developed method can be further extended to other districting problems under balancing and connectivity constraints.     

Scheduling jobs with ready times and precedence constraints on parallel batch machines using metaheuristics

In this paper, we discuss a scheduling problem for parallel batch machines where the jobs have ready times. Problems of this type can be found in semiconductor wafer fabrication facilities (wafer fabs). In addition, we consider precedence constraints among the jobs. Such constraints arise, for example, in scheduling subproblems of the shifting bottleneck heuristic when complex job shop scheduling problems are tackled. We use the total weighted tardiness as the performance measure to be optimized. Hence, the problem is NP-hard and we have to rely on heuristic solution approaches. We consider a variable neighborhood search (VNS) scheme and a greedy randomized adaptive search procedure (GRASP) to compute efficient solutions. We assess the performance of the two metaheuristics based on a large set of randomly generated problem instances and based on instances from the literature. The obtained computational results demonstrate that VNS is a very fast heuristic that quickly leads to high-quality solutions, whereas the GRASP is slightly outperformed by the VNS approach. However, the GRASP approach has the advantage that it can be parallelized in a more natural manner compared to VNS.     

A hybrid column generation with GRASP and path relinking for the network load balancing problem

In this paper, a hybrid meta-heuristic is proposed which combines the GRASP with path relinking method and Column Generation. The key idea of this method is to run a GRASP with path relinking search on a restricted search space, defined by Column Generation, instead of running the search on the complete search space of the problem. Moreover, column generation is used not only to compute the initial restricted search space but also to modify it during the whole algorithm. The proposed heuristic is used to solve the network load balancing problem: given a capacitated telecommunications network with single path routing and an estimated traffic demand matrix, the network load balancing problem is the determination of a routing path for each traffic commodity such that the network load balancing is optimized, i.e., the worst link load is minimized, among all such solutions, the second worst link load is minimized, and continuing in this way until all link loads are minimized. The computational results presented in this paper show that, for the network load balancing problem, the proposed heuristic is effective in obtaining better quality solutions in shorter running times.     

Hybridising GRASP and network flows in the solution of a medical school scheduling problem

This paper tackles the problem of allocating medical students to clinical specialities over a number of time periods. Each speciality is offered by a number of consultant led ??firms?? and the objective is to optimise the schedule in terms of ensuring a broad range of experience for each student, whilst ensuring that every student covers every speciality exactly once without exceeding the number of places available in each firm. The balance between feasibility and optimality is a key issue. We develop a family of GRASP heuristics for the problem, all based on the same local search heuristic, but using a variety of constructions. These use different building blocks, different score functions, and different ways of balancing feasibility and optimality. Empirical tests show that the best heuristic, based on large building blocks facilitated by the use of a network flow model, and enhanced by feedback in the form of partial reconstructions, performs extremely well on real data, and is able to produce acceptable solutions on more challenging artificial problem instances.     

A scatter search approach for the minimum sum-of-squares clustering problem

A metaheuristic procedure based on the scatter search approach is proposed for the non-hierarchical clustering problem under the criterion of minimum sum-of-squares clustering. This algorithm incorporates procedures based on different strategies, such as local search, GRASP, tabu search or path relinking. The aim is to obtain quality solutions with short computation times. A series of computational experiments has been performed. The proposed algorithm obtains better results than previously reported methods, especially with small numbers of clusters.     

GRASP and path relinking for the equitable dispersion problem

The equitable dispersion problem consists in selecting a subset of elements from a given set in such a way that a measure of dispersion is maximized. In particular, we target the Max-Mean dispersion model in which the average distance between the selected elements is maximized. We first review previous methods and mathematical formulations for this and related dispersion problems and then propose a GRASP with a Path Relinking in which the local search is based on the Variable Neighborhood methodology. Our method is specially suited for instances in which the distances represent affinity and are not restricted to take non-negative values. The computational experience with 120 instances shows the merit of the proposed procedures compared to previous methods.     

An annotated bibliography of GRASP – Part I: Algorithms

A greedy randomized adaptive search procedure (GRASP) is a metaheuristic for combinatorial optimization. It is a multi-start or iterative process, in which each GRASP iteration consists of two phases, a construction phase, in which a feasible solution is produced, and a local search phase, in which a local optimum in the neighborhood of the constructed solution is sought. Since 1989, numerous papers on the basic aspects of GRASP, as well as enhancements to the basic metaheuristic have appeared in the literature. GRASP has been applied to a wide range of combinatorial optimization problems, ranging from scheduling and routing to drawing and turbine balancing. This is the first of two papers with an annotated bibliography of the GRASP literature from 1989 to 2008. This paper covers algorithmic aspects of GRASP.     

A multi-objective GRASP for partial classification

Metaheuristic algorithms have been used successfully in a number of data mining contexts and specifically in the production of classification rules. Classification rules describe a class of interest or a subset of this class, and as such may also be used as an aid in prediction. The production and selection of classification rules for a particular class of the database is often referred to as partial classification. Since partial classification rules are often evaluated according to a number of conflicting objectives, the generation of such rules is a task that is well suited to a multi-objective (MO) metaheuristic approach. In this paper we discuss how to adapt well known MO algorithms for the task of partial classification. Additionally, we introduce a new MO algorithm for this task based on a greedy randomized adaptive search procedure (GRASP). GRASP has been applied to a number of problems in combinatorial optimization, but it has very seldom been used in a MO setting, and generally only through repeated optimization of single objective problems, using either linear combinations of the objectives or additional constraints. The approach presented takes advantage of some specific characteristics of the data mining problem being solved, allowing for the very effective construction of a set of solutions that form the starting point for the local search phase of the GRASP. The resulting algorithm is guided solely by the concepts of dominance and Pareto-optimality. We present experimental results for our partial classification GRASP and other MO metaheuristics. These show that such algorithms are generally very well suited to this data mining task and furthermore, the GRASP brings additional efficiency to the search for partial classification rules.     

Hybrid ensemble selection algorithm incorporating GRASP with path relinking

The greedy randomized adaptive search procedure (GRASP) is an iterative two-phase multi-start metaheuristic procedure for a combination optimization problem, while path relinking is an intensification procedure applied to the solutions generated by GRASP. In this paper, a hybrid ensemble selection algorithm incorporating GRASP with path relinking (PRelinkGraspEnS) is proposed for credit scoring. The base learner of the proposed method is an extreme learning machine (ELM). Bootstrap aggregation (bagging) is used to produce multiple diversified ELMs, while GRASP with path relinking is the approach for ensemble selection. The advantages of the ELM are inherited by the new algorithm, including fast learning speed, good generalization performance, and easy implementation. The PRelinkGraspEnS algorithm is able to escape from local optima and realizes a multi-start search. By incorporating path relinking into GRASP and using it as the ensemble selection method for the PRelinkGraspEnS the proposed algorithm becomes a procedure with a memory and high convergence speed. Three credit datasets are used to verify the efficiency of our proposed PRelinkGraspEnS algorithm. Experimental results demonstrate that PRelinkGraspEnS achieves significantly better generalization performance than the classical directed hill climbing ensemble pruning algorithm, support vector machines, multi-layer perceptrons, and a baseline method, the best single model. The experimental results further illustrate that by decreasing the average time needed to find a good-quality subensemble for the credit scoring problem, GRASP with path relinking outperforms pure GRASP (i.e., without path relinking).     

Heuristics for the capacitated dispersion problem

In this paper, we investigate the adaptation of the greedy randomized adaptive search procedure (GRASP) and variable neighborhood descent (VND) methodologies to the capacitated dispersion problem. Dispersion and diversity problems arise in the placement of undesirable facilities, workforce management, and social media, among others. Maximizing diversity deals with selecting a subset of elements from a given set in such a way that the distance among the selected elements is maximized. We target here a realistic variant with capacity constraints for which a heuristic with a performance guarantee was previously introduced. In particular, we propose a hybridization of GRASP and VND implementing within the strategic oscillation framework. To evaluate the performance of our heuristic, we perform extensive experimentation to first set key search parameters, and then compare the final method with the previous heuristic. Additionally, we propose a mathematical model to obtain optimal solutions for small‐sized instances, and compare our solutions with the well‐known LocalSolver software.     

The fixed set search applied to the power dominating set problem

In this article, we focus on solving the power dominating set problem and its connected version. These problems are frequently used for finding optimal placements of phasor measurement units in power systems. We present an improved integer linear program (ILP) for both problems. In addition, a greedy constructive algorithm and a local search are developed. A greedy randomised adaptive search procedure (GRASP) algorithm is created to find near optimal solutions for large scale problem instances. The performance of the GRASP is further enhanced by extending it to the novel fixed set search (FSS) metaheuristic. Our computational results show that the proposed ILP has a significantly lower computational cost than existing ILPs for both versions of the problem. The proposed FSS algorithm manages to find all the optimal solutions that have been acquired using the ILP. In the last group of tests, it is shown that the FSS can significantly outperform the GRASP in both solution quality and computational cost.     

GRASP with evolutionary path-relinking for the capacitated arc routing problem

The Capacitated Arc Routing Problem (CARP) is a well-known NP-hard combinatorial optimization problem where, given an undirected graph, the objective is to find a minimum cost set of tours servicing a subset of required edges under vehicle capacity constraints. There are numerous applications for the CARP, such as street sweeping, garbage collection, mail delivery, school bus routing, and meter reading. A Greedy Randomized Adaptive Search Procedure (GRASP) with Path-Relinking (PR) is proposed and compared with other successful CARP metaheuristics. Some features of this GRASP with PR are (i) reactive parameter tuning, where the parameter value is stochastically selected biased in favor of those values which historically produced the best solutions in average; (ii) a statistical filter, which discard initial solutions if they are unlikely to improve the incumbent best solution; (iii) infeasible local search, where high-quality solutions, though infeasible, are used to explore the feasible/infeasible boundaries of the solution space; (iv) evolutionary PR, a recent trend where the pool of elite solutions is progressively improved by successive relinking of pairs of elite solutions. Computational tests were conducted using a set of 81 instances, and results reveal that the GRASP is very competitive, achieving the best overall deviation from lower bounds and the highest number of best solutions found.     

An annotated bibliography of GRASP–Part II: Applications

A greedy randomized adaptive search procedure (GRASP) is a metaheuristic for combinatorial optimization. It is a multi-start or iterative process, in which each GRASP iteration consists of two phases, a construction phase, in which a feasible solution is produced, and a local search phase, in which a local optimum in the neighborhood of the constructed solution is sought. Since 1989, numerous papers on the basic aspects of GRASP, as well as enhancements to the basic metaheuristic, have appeared in the literature. GRASP has been applied to a wide range of combinatorial optimization problems, ranging from scheduling and routing to drawing and turbine balancing. This is the second of two papers with an annotated bibliography of the GRASP literature from 1989 to 2008. In the companion paper, algorithmic aspects of GRASP are surveyed. In this paper, we cover the literature where GRASP is applied to scheduling, routing, logic, partitioning, location, graph theory, assignment, manufacturing, transportation, telecommunications, biology and related fields, automatic drawing, power systems, and VLSI design.     

Local search for load balancing problems for servers with large dimension

We consider a new load balancing model that arises in the processing of user requests for files located on a given set of servers. The optimization criterion is the total excess of actual load over the limit load. In order to redistribute the load and minimize the criterion, files can be moved between the servers. We show that if there are no other constraints related to the stage of moving the files, then this problem is equivalent to a problem previously considered in literature. For this special case of this problem, we propose a stochastic local search scheme that combines a special procedure for fast querying of the neighborhoods and a procedure of non-aggravating modification of intermediate solutions. Results of numerical experiments show that the proposed approach is able to find high-quality solutions for instances of large dimension under tight time constraints.     

A GRASP approach to the container-loading problem

The container-loading problem aims to determine the arrangement of items in a container. We present GRMODGRASP, a new algorithm for the CLP based on the GRASP (greedy randomized adaptive search procedure) paradigm. We evaluate GRMODGRASP'S performance in terms of volume use and load stability and by comparing it with nine well-known algorithms. Our approach produces solutions that surpass other approaches' solutions in terms of volume use and cargo stability.     

TSP问题的一种改进的GRASP算法

本文对Marinakis等提出的扩展邻域GRASP算法进行改进。首先使用最近α值方法构造初始TSP回路,然后运用混合的局部搜索即2-opt算法、双桥策略和3-opt算法来改进初始回路,并且引进α-nearness候选集和don’t-lookbit技术来提高搜索速度。实验结果表明,本文提出的GRASP能够在合理的时间内得到很好的解,并且解的质量优于M~rinakis等提出的扩展邻域GRASP算法得到的解。     

Multiple Phase Neighborhood Search-GRASP for the Capacitated Vehicle Routing Problem

Greedy Randomized Adaptive Search Procedure (GRASP) has been proved to be a very efficient algorithm for the solution of the Traveling Salesman Problem. Also, it has been proved that expanding the local search with the use of two or more different local search strategies helps the algorithm to avoid trapping in a local optimum. In this paper, a new modified version of GRASP, called Multiple Phase Neighborhood Search-GRASP (MPNS-GRASP), for the solution of the Vehicle Routing Problem is proposed. In this method, a stopping criterion based on Lagrangean Relaxation and Subgradient Optimization is utilized. In addition, a different way for expanding the neighborhood search is used based on a new strategy, the Circle Restricted Local Search Moves strategy. The algorithm was tested on two sets of benchmark instances and gave very satisfactory results. In both sets of instances the results have solution qualities with average values near to the optimum values and in a number of them the algorithm finds the optimum. The computational time of the algorithm is decreased significantly compared to other heuristic and metaheuristic algorithms due to the fact that the new strategy, the Expanding Neighborhood Search Strategy, is used.     

求解多车型校车路径问题的混合集合划分的GRASP算法

针对不同规划场景下具有不同优化目标的多车型校车路径问题(HSBRP),提出一种混合集合划分(SP)的贪婪随机自适应(Greedy Randomized Adaptive Search Procedure,GRASP)算法。根据GRASP算法寻优过程中产生的路径信息构建SP模型,然后使用CPLEX精确优化器对SP模型进行求解。为了适应不同类型的HSBRP问题,改进GRASP的初始解构造函数得到一个可行解,并将其对应的路径放入路径池;在局部搜索过程中应用多种邻域结构和可变邻域下降(VND)来提升解的质量,同时在路径池中记录在搜索过程中得到提升的路径和在每次迭代中得到局部最好解的路径信息。使用基准测试案例进行测试,实验结果表明在GRASP算法中,混合SP能够有效地提高算法的求解性能和稳定性,并且该算法能适应不同优化目标下车型混合和车辆数限制两类HSBRP的求解;与现有算法的比较结果再次验证了所提算法的有效性。     

A GRASP based solution approach to solve cardinality constrained portfolio optimization problems

In the current work, a solution methodology which combines a meta-heuristic algorithm with an exact solution approach is presented to solve cardinality constrained portfolio optimization (CCPO) problem. The proposed method is comprised of two levels, namely, stock selection and proportion determination. In stock selection level, a greedy randomized adaptive search procedure (GRASP) is developed. Once the stocks are selected the problem reduces to a quadratic programming problem. As GRASP ensures cardinality constraints by selecting predetermined number of stocks and quadratic programming model ensures the remaining problem constraints, no further constraint handling procedures are required. On the other hand, as the problem is decomposed into two sub-problems, total computational burden on the algorithm is considerably reduced. Furthermore, the performance of the proposed algorithm is evaluated by using benchmark data sets available in the OR Library. Computational results reveal that the proposed algorithm is competitive with the state of the art algorithms in the related literature.