Catálogo de publicaciones - libros

The resource-constrained project scheduling problem (RCP- SP) is one of the most challenging problems in project scheduling. During the last couple of years many heuristic procedures have been developed for this problem, but still these procedures often fail in finding near-optimal solutions for more challenging problem instances. In this paper, we present a new genetic algorithm (GA) that, in contrast of a conventional GA, makes use of two separate populations. This bi-population genetic algorithm (BPGA) operates on both a population of left-justified schedules and a population of right-justified schedules in order to fully exploit the features of the iterative forward/backward scheduling technique. Comparative computational results reveal that this procedure can be considered as today’s best performing RCPSP heuristic.

Assume that the system of two polynomial equations (,) = 0 and (,) = 0 has a finite number of solutions. Then the solution consists of pairs of an -value and an -value. In some cases conventional methods to calculate these solutions give incorrect results and are complicated to implement due to possible degeneracies and multiple roots in intermediate results. We propose and test a two-step method to avoid these complications. First all -roots and all -roots are calculated independently. Taking the multiplicity of the roots into account, the number of -roots equals the number of -roots. In the second step the -roots and -roots are matched by constructing a weighted bipartite graph, where the -roots and the -roots are the nodes of the graph, and the errors are the weights. Of this graph the minimum weight perfect matching is computed. By using a multidimensional matching method this principle may be generalized to more than two equations.

The vehicle routing problem (VRP) has been extensively studied because of the interest in its application in logistics and supply chain management. In this paper, we develop a hybrid algorithm (IACS-SA) that combines the strengths of improved ant colony system (IACS) and simulated annealing (SA) algorithm. The results of computational experiments on fourteen VRP benchmark problems show that our IACS-SA produces better solutions than those of other ACS in the literature. The results also indicate that such a hybrid algorithm is comparable with other meta-heuristic algorithms.

This paper proposes a coevolutionary classification method to discover classifiers for multidimensional pattern classification problems with continuous features. The classification problems may be decomposed into two sub-problems, which are feature selection and classifier adaptation. A coevolutionary classification method is designed by coordinating the two sub-problems, whose performances are affected by each other. The proposed method establishes a group of partial sub-regions, defined by regional feature set, and then fits a finite number of classifiers to the data pattern by combining a genetic algorithm and a local adaptation algorithm in every sub-region. A cycle of the cooperation loop is completed by evolving the sub-regions based on the evaluation results of the fitted classifiers located in the corresponding sub-regions. The classifier system has been tested with well-known data sets from the UCI machine-learning database, showing superior performance to other methods such as the nearest neighbor, decision tree, and neural networks.

We present closed form analytical results to show the throughput performance of a discrete-window closed-loop conveyors system serving a user-specified set of stations with intermixed load/unload stations. The buffer capacity at the unloading stations is finite; loads that encounter a full buffer (i.e., blocked loads) are assumed to recirculate around the loop to try again. Given the job flow and routing data as well as the configuration of the conveyor loop, we present an analytical approach to approximate the expected overflow of loads on the conveyor (due to blocked loads). Given the expected overflow, we also show the stability condition for the conveyor system.

This paper explores the problem of making ordering decisions for multi-items with various constraints on the supply chain. With the advent of supply chain management, depots or warehouses fulfill a strategic role of achieving the logistics objectives of shorter cycle times, lower inventories, lower costs and better customer service. Many companies consider both their cost effectiveness and market proficiency to depend primarily on efficient logistics management. Depot management presently is considered a key to strengthening company logistics. In this paper, we propose a novel multi-items ordering model with the mixed parts transportation problem based on the depot system in a supply chain. Order range is especially introduced and used as decision parameters instead of order point to order multi-items. Finally, we test the model with a numerical example and show computational results that verify the effectiveness of the proposed model.

This paper addresses a method to obtain the best sequence for no-wait flowshop scheduling with precedence constraints and sequence dependent setup time. The system is made up of a set of machines of various types and there is no interruption between tasks in a job. The objective is to determine the job sequence for processing with minimum makespan. The sequencing problem with precedence constraints and sequence dependent setup times is equivalent to the traveling salesman problem. A mixed integer programming (MIP) model is presented to obtain the best schedule. It is known that the MIP model for no-wait flowshop scheduling is NP-hard. Heuristic algorithms to gain the best job sequence are developed. From the experiments, it is found that the proposed algorithm generates the best job sequence efficiently.

The minimization of nonlinearly constrained network flow problems can be performed by exploiting the efficiency of the network flow techniques. It lies in minimizing approximately a series of (augmented) Lagrangian functions including only the side constraints, subject to balance constraints in the nodes and capacity bounds. One of the drawbacks of the multiplier methods with quadratic penalty function when is applied to problems with inequality constraints is that the corresponding augmented Lagrangian function is not twice continuously differentiable even if the cost and constraint functions are. The author’s purpose is to put forward two methods that overcome this difficulty: the exponential multiplier method and the -subgradient method, and to compare their efficiency with that of the quadratic multiplier method and that of the codes MINOS and LOQO. The results are encouraging.

This article presents an algorithm that finds an -feasible solution relatively to some constraints of a linear program. The algorithm is a first-order feasible directions method with constant stepsize that attempts to find the minimizer of an exponential penalty function. When embedded with bisection search, the algorithm allows for the approximated solution of linear programs. We present applications of this framework to set-partitioning problems and report some computational results with first-order and second-order implementations.

This paper develops a METRIC-like method to optimize inventory allocations for low-demand items in a multi-echelon finished goods distribution system. The allocations are made by trading-off total inventory investment against service levels as measured by the stockout probabilities. We focus on a one-product three-echelon system and develop a solution procedure for such a system. We then apply that model and solution procedure to an illustrative example and compare the resulting inventory policies to optimal policies for a two-echelon system.