Home  About us  Contact
 

Computational Experience (computational + experience)

Distribution by Scientific Domains
Mathematics and Statistics60%

Selected Abstracts

Scheduling dispensing and counting in secondary pharmaceutical manufacturing

AICHE JOURNAL, Issue 5 2009
Michele Ciavotta
Abstract In this article, we describe a general methodology for operations scheduling in dispensing and counting departments of pharmaceutical manufacturing plants. The departments are modeled as a multiobjective parallel machines scheduling problem under a number of both standard and realistic constraints, such as release times, due dates and deadlines, particular sequence-dependent setup times, machine unavailabilities, and maximum campaign size. Main characteristics of the methodology are the modularity of the solution algorithms, the adaptability to different objectives and constraints to fulfill production requirements, the easiness of implementation, and the ability of incorporating human experience in the scheduling algorithms. Computational experience carried out on two case studies from a real pharmaceutical plant shows the effectiveness of this approach. © 2009 American Institute of Chemical Engineers AIChE J, 2009 [source]

A supply chain design model with unreliable supply

NAVAL RESEARCH LOGISTICS: AN INTERNATIONAL JOURNAL, Issue 8 2007
Lian Qi
Abstract Uncertainties abound within a supply chain and have big impacts on its performance. We propose an integrated model for a three-tiered supply chain network with one supplier, one or more facilities and retailers. This model takes into consideration the unreliable aspects of a supply chain. The properties of the optimal solution to the model are analyzed to reveal the impacts of supply uncertainty on supply chain design decisions. We also propose a general solution algorithm for this model. Computational experience is presented and discussed. © 2007 Wiley Periodicals, Inc. Naval Research Logistics, 2007 [source]

A genetic algorithm approach to solving the anti-covering location problem

EXPERT SYSTEMS, Issue 5 2006
Sohail S. Chaudhry
Abstract: In this paper we address the problem of locating a maximum weighted number of facilities such that no two are within a specified distance from each other. A natural process of evolution approach, more specifically a genetic algorithm, is proposed to solve this problem. It is shown that through the use of a commercially available spreadsheet-based genetic algorithm software package, the decision-maker with a fundamental knowledge of spreadsheets can easily set up and solve this optimization problem. Also, we report on our extensive computational experience using three different data sets. [source]

Information security: Designing a stochastic-network for throughput and reliability,,

NAVAL RESEARCH LOGISTICS: AN INTERNATIONAL JOURNAL, Issue 7 2009
Jeffrey Schavland
Abstract Todas information and communication network requires a design that is secure to tampering. Traditional performance measures of reliability and throughput must be supplemented with measures of security. Recognition of an adversary who can inflict damage leads toward a game-theoretic model. Through such a formulation, guidelines for network designs and improvements are derived. We opt for a design that is most robust to withstand both natural degradation and adversarial attacks. Extensive computational experience with such a model suggests that a Nash-equilibrium design exists that can withstand the worst possible damage. Most important, the equilibrium is value-free in that it is stable irrespective of the unit costs associated with reliability vs. capacity improvement and how one wishes to trade between throughput and reliability. This finding helps to pinpoint the most critical components in network design. From a policy standpoint, the model also allows the monetary value of information-security to be imputed. © 2009 Wiley Periodicals, Inc. Naval Research Logistics, 2009 [source]

The 2-path network problem

NETWORKS: AN INTERNATIONAL JOURNAL, Issue 3 2004
Geir Dahl
Abstract Given a graph with nonnegative edge weights and a set D of node pairs, the 2- path network problem requires a minimum weight set of edges such that the induced subgraph contains a path with one or two edges connecting each pair in D. The problem is NP -hard. We present two integer programming models for the problem and study properties of associated polytopes, including cutting planes. Two approximation algorithms are suggested and analyzed. Some computational experience is reported. © 2004 Wiley Periodicals, Inc. [source]