Clara Novoa

An approximate dynamic programming approach for the vehicle routing problem with stochastic demands

This paper examines approximate dynamic programming algorithms for the single-vehicle routing problem with stochastic demands from a dynamic or reoptimization perspective. The methods extend the rollout algorithm by implementing different base sequences (i.e. a priori solutions), look-ahead policies, and pruning schemes. The paper also considers computing the cost-to-go with Monte Carlo simulation in addition to direct approaches. The best new method found is a two-step lookahead rollout started with a stochastic base sequence. The routing cost is about 4.8% less than the one-step rollout algorithm started with a deterministic sequence. Results also show that Monte Carlo cost-to-go estimation reduces computation time 65% in large instances with little or no loss in solution quality. Moreover, the paper compares results to the perfect information case from solving exact a posteriori solutions for sampled vehicle routing problems. The confidence interval for the overall mean difference is (3.56%, 4.11%).

Optimising the automated plasma cutting process by design of experiments

Automated plasma cutting is an effective process for building complex, two-dimensional metallic parts in a short period of time. Because the plasma cutting machine has several factors or input variables to control (e.g., current, cutting speed, torch height, etc.) and the process requires compliance with a variety of part quality characteristics or response variables (e.g., flatness, clean cut, bevel angle, etc.), it is difficult to find a machine setting that improves the overall quality of the manufactured parts. This research was conducted to discover the relevant factors that affect the parts surface quality characteristics and the optimum machine settings by implementing a design of experiments and following a response surface methodology approach. Desirability functions were used to optimise the automated plasma cutting process. Final results identified an optimal machine configuration that facilitates the fabrication of parts with close-to-perfect quality for all the 18 quality responses considered.

Bootstrap methods for analyzing time studies and input data for simulations

Purpose – The purpose of this paper is to present bootstrapping as an alternative statistical methodology to analyze time studies and input data for discrete‐event simulations. Bootstrapping is a non‐parametric technique to estimate the sampling distribution of a statistic by doing repeated sampling (i.e. resampling) with replacement from an original sample. This paper proposes a relatively simple implementation of bootstrap techniques to time study analysis.Design/methodology/approach – Using an inductive approach, this work selects a typical situation to conduct a time study, applies two bootstrap procedures for the statistical analysis, compares bootstrap to traditional parametric approaches, and extrapolates general advantages of bootstrapping over parametric approaches.Findings – Bootstrap produces accurate inferences when compared to those from parametric methods, and it is an alternative when the underlying parametric assumptions are not met.Research limitations/implications – Research results cont...

A SIMD tabu search implementation for solving the quadratic assignment problem with GPU acceleration

In the Quadratic Assignment Problem (QAP), n units (usually departments, machines, or electronic components) must be assigned to n locations given the distance between the locations and the flow between the units. The goal is to find the assignment that minimizes the sum of the products of distance traveled and flow between units. The QAP is a combinatorial problem difficult to solve to optimality even for problems where n is relatively small (e.g., n = 30). In this paper, we develop a parallel tabu search algorithm to solve the QAP and leverage the compute capabilities of current GPUs. The single instruction multiple data (SIMD) algorithm is implemented on the Stampede cluster hosted by the Texas Advanced Computing Center (TACC) at the University of Texas at Austin. We enhance our implementation by exploiting the dynamic parallelism made available in the Nvidia Kepler high performance computing architecture. On a series of experiments on the well-known QAPLIB data sets, our algorithm produces solutions that are as good as the best known ones posted in QAPLIB. The worst case percentage of accuracy we obtained was 0.83%. Given the applicability of QAP, our algorithm has very good potential to accelerate scholarly research in Engineering, in the fields of Operations Research and design of electronic devices. To the best of our knowledge, this work is the first to successfully parallelize the tabu search metaheuristic to solve the QAP with the recency-based feature, implemented serially in [10]. Our work is also the first to exploit GPU dynamic parallelism in a tabu search metaheuristic to solve the QAP.

A SIMD Solution for the Quadratic Assignment Problem with GPU Acceleration

In the Quadratic Assignment Problem (QAP), n units (usually departments, machines, or electronic components) must be assigned to n locations given the distance between the locations and the flow between the units. The goal is to find the assignment that minimizes the sum of the products of distance traveled and flow between units. The QAP is a combinatorial problem difficult to solve to optimality even for problems where n is relatively small (e.g., n = 30). In this paper, we solve the QAP problem using a parallel algorithm that employs a 2-opt heuristic and leverages the compute capabilities of current GPUs. The algorithm is implemented on the Stampede cluster hosted by the Texas Advanced Computing Center (TACC) at the University of Texas at Austin and on a GPU-equipped server housed at Texas State University. We enhance our implementation by fine tuning the occupancy levels and by exploiting inter-thread data locality through improved shared memory allocation. On a series of experiments on the well-known QAPLIB data sets, our algorithm, on average, outperforms an OpenMP implementation by a factor of 16.31 and a Tabu search based GPU implementation by a factor of 58.61. Given the wide applicability of QAP, the algorithm we propose has very good potential to accelerate the discovery in scholarly research in Engineering, particularly in the fields of Operations Research and design of electronic devices.

Managing performance based logistics by balancing reliability and spare parts stocking

This paper investigates a quantitative approach for performance based logistics management in the presence of uncertainty. Our study concentrates on the situation where the customer purchased the capital equipment from the original equipment manufacturer who also provides the after-sales service. The service provider will be rewarded monetarily if equipment availability is kept above a pre-specified level. We derived an analytical model to characterize the equipment availability incorporating five performance drivers. This analytical insight into the equipment availability allows us to assess the trade-off between product reliability and the spare parts level. The equipment availability metric is examined under various scenarios by varying the usage rate, repair turn-around time, and the availability target.

Supplier Selection and Order Allocation Based on Integer Programming

The ability to assess and select new suppliers quickly and efficiently is a critical requirement for improving the agility of manufacturing supply chains. The Digital Manufacturing Market DMM is a web-based platform for intelligent supply chain configuration. This research enhances the DMMs performance by developing a column generation method for solving the supplier selection problem. The objective of the proposed method is to maximize the technological competencies of the selected suppliers while meeting their capacity constraints. The column generation method resolves the issue of limited scalability of a traditional linear programming formulation and can be integrated into the DMM. Additionally, using test generated problems, this research evaluates the effect on reducing the threshold distance traveled by semi-finished parts in the work orders. The results show that an economy of distance can be imposed with little effect on average match compatibility.

On the Distribution of the Usual Estimator of C and Some Applications in SPC

ABSTRACT This article presents a new and appealing derivation of the distribution of Ĉ pk when the quality characteristic is normally distributed. Our derivation is based on the folded normal distribution and the fact that sample mean and sample standard deviation are independent. The final result is a complete and self-contained formula that can be easily implemented in any modern math package. We illustrate the use of the pdf of Ĉ pk in joint hypothesis testing of process parameters and in process control and monitoring. The ARL performance of a proposed control chart based on Ĉ pk is compared against the performance of and R charts.

A zero-carbon supply chain model: minimising levelised cost of onsite renewable generation

ABSTRACT This paper develops a linear optimisation model to investigate whether it is feasible to operate a net-zero-carbon supply chain network via 100% of onsite wind and solar generation. In particular, this work determines the technology, location and size of renewable generating units in an integrated production–transportation system with the goal of minimising the levelised cost of renewable energy. Numerical experiments show that it is technically feasible and economically viable to realise a zero-carbon supply chain operation provided the local facility possesses a medium wind speed and the installation cost of photovoltaics drops to

Autotuning GPU-Accelerated QAP Solvers for Power and Performance

2 per watt.