Yuehwern Yih

Scheduling elective surgery under uncertainty and downstream capacity constraints

The objective of this study is to generate an optimal surgery schedule of elective surgery patients with uncertain surgery operations, which includes uncertainty in surgery durations and the availability of downstream resources such as surgical intensive care unit (SICU) over multi-periods. The stochastic optimization is adapted and the sample average approximation (SAA) method is proposed for obtaining an optimal surgery schedule with respect to minimizing the total cost of patient costs and overtime costs. A computational experiment is presented to evaluate the performance of the proposed method.

A Neural Network Integrated Decision Support System for Condition-Based Optimal Predictive Maintenance Policy

This paper develops an integrated neural-network-based decision support system for predictive maintenance of rotational equipment. The integrated system is platform-independent and is aimed at minimizing expected cost per unit operational time. The proposed system consists of three components. The first component develops a vibration-based degradation database through condition monitoring of rolling element bearings. In the second component, an artificial neural network model is developed to estimate the life percentile and failure times of roller bearings. This is then used to construct a marginal distribution. The third component consists of the construction of a cost matrix and probabilistic replacement model that optimizes the expected cost per unit time. Furthermore, the integrated system consists of a heuristic managerial decision rule for different scenarios of predictive and corrective cost compositions. Finally, the proposed system can be applied in various industries and different kinds of equipment that possess well-defined degradation characteristics

An elective surgery scheduling problem considering patient priority

This paper addresses a scheduling problem where patients with different priorities are scheduled for elective surgery in a surgical facility, which has a limited capacity. When the capacity is available, patients with a higher priority are selected from the waiting list and put on the schedule. At the beginning of each period, a decision of the number of patients to be scheduled is made based on the trade-offs between the cost for overtime work and the cost for surgery postponement. A stochastic dynamic programming model is formulated to address this problem. A structural analysis of the proposed model is conducted to understand the properties of an optimal schedule policy. Based on the structural analysis, bounds on feasible actions are incorporated into a value iteration algorithm, and a brief computation experiment shows the improvement in computational efficiency. Numerical examples show that the consideration of patient priority results in significant differences in surgery schedules from the schedule that ignores the patient priority.

Integrating nutrition support for food-insecure patients and their dependents into an HIV care and treatment program in western Kenya.

The Academic Model Providing Access to Healthcare (AMPATH) is a partnership between Moi Teaching and Referral Hospital, Moi University School of Medicine, and a consortium of universities led by Indiana University. AMPATH has over 50,000 patients in active care in 17 main clinics around western Kenya. Despite antiretroviral therapy, many patients were not recovering their health because of food insecurity. AMPATH therefore established partnerships with the World Food Program and United States Agency for International Development and began high-production farms to complement food support. Today, nutritionists assess all AMPATH patients and dependents for food security and refer those in need to the food program. We describe the implementation, challenges, and successes of this program.

Generic kanban systems for dynamic environments

Abstract Kanbans have shown successful results in lowering inventory and shortening lead time in repetitive production systems. Unfortunately, such systems are not applicable to production environments with dynamic characteristics. Here a modified kanban system, the generic kanban system, is proposed for such dynamic environments. The generic kanban system behaves similarly to the push system except that it is more flexible with respect to system performance and more robust as to the location of the bottleneck. The simulation results that the generic kanban system is dominant over the dedicated kanban system and the CONWIP system. The adaptability of such a system to dynamic environments is justified.

Optimal linear combinations of neural networks: an overview

Neural networks based modeling often involves trying multiple networks with different architectures and/or training parameters in order to achieve acceptable model accuracy. Typically, one of the trained NNs is chosen as best, while the rest are discarded. Hashem and Schmeiser (1992) propose using optimal linear combinations of a number of trained neural networks instead of using a single best network. In this paper, we discuss and extend the idea of optimal linear combinations of neural networks. Optimal linear combinations are constructed by forming weighted sums of the corresponding outputs of the networks. The combination-weights are selected to minimize the mean squared error with respect to the distribution of random inputs. Combining the trained networks may help integrate the knowledge acquired by the component networks and thus improve model accuracy. We investigate some issues concerning the estimation of the optimal combination-weights and the role of the optimal linear combination in improving model accuracy for both well-trained and poorly trained component networks. Experimental results based on simulated data are included. For our examples, the model accuracy resulting from using estimated optimal linear combinations is better than that of the best trained network and that of the simple averaging of the outputs of the component networks.<<ETX>>

An agent-based production control framework for multiple-line collaborative manufacturing

This research focuses on constructing an agent-based collaborative production control framework that is capable of conducting scheduling and dispatching functions among production entities, as well as within them, in a collaborative manufacturing environment. The proposed framework utilizes autonomous agent and weighted functions for distributed decision-making while all agents work in active and collaborative ways to help each other make decisions. This collaborative control framework is capable of realizing and seeking balances among heterogeneous objectives of the production entities within a collaborative manufacturing system. The agents in this control framework were constructed with an object-oriented prospective so that a production entity can join or depart from the control scheme without affecting the rest of the framework. Simple index values, instead of detailed data, were used for information exchange among agents. This can greatly reduce the communication and computation load of the control system and keep detailed production information confidential while the agents in the system could belong to different companies. In this research we created a simulation model of a real-world multi-line elevator manufacturing system as the test bed to evaluate the performance of the proposed control framework. Two other control strategies with different levels of collaboration were applied to the simulation model to compare and evaluate the performances of the proposed control strategy. Results of the simulation show that multiple objectives of the production entities can be realized.

INTEGRATION OF INDUCTIVE LEARNING AND NEURAL NETWORKS FOR MULTI-OBJECTIVE FMS SCHEDULING

In this paper, we propose an integrated approach of inductive learning and competitive neural networks for developing multi-objective flexible manufacturing system (FMS) schedulers. Simulation and competitive neural networks are applied sequentially to extract a set of classified training data which is used to create a compact set of scheduling rules through inductive learning. The FMS scheduler can assist the operator to make decisions in real time, while satisfying multiple objectives desired by the operator. A simulation-based experiment is performed to evaluate the performance of the resulting scheduler.

An algorithm for hoist scheduling problems

This paper proposes an algorithm for hoist scheduling problems in a flexible PCB electroplating line where there is no buffer among workstations. Due to its chemical process nature, the processing times have to be controlled within a specified range (i.e. between a maximum processing time and a minimum processing time). The product will become defective if this constant is violated. The objective is to maximize throughput with no defective product. The proposed algorithm schedules jobs entering the system based on the specified range of processing time. A simulation study shows that the proposed algorithm outperforms the basic algorithm that schedules jobs based on the minimum processing times.

Knowledge acquisition through information granulation for imbalanced data

Abstract When learning from imbalanced/skewed data, which almost all the instances are labeled as one class while far few instances are labeled as the other class, traditional machine learning algorithms tend to produce high accuracy over the majority class but poor predictive accuracy over the minority class. This paper proposes a novel method called ‘knowledge acquisition via information granulation’ (KAIG) model which not only can remove some unnecessary details and provide a better insight into the essence of data but also effectively solve ‘class imbalance’ problems. In this model, the homogeneity index (H-index) and the undistinguishable ratio (U-ratio) are successfully introduced to determine a suitable level of granularity. We also developed the concept of sub-attributes to describe granules and tackle the overlapping among granules. Seven data sets from UCI data bank, including one imbalanced diagnosis data (pima-Indians-diabetes), are provided to evaluate the effectiveness of KAIG model. By using different performance indexes, overall accuracy, G-mean and Receiver Operation Characteristic (ROC) curve, the experimental results comparing with C4.5 and Support Vector Machine (SVM) demonstrate the superiority of our method.