This paper examines a decentralized admission control system with partial capacity sharing in a hospital setting. The admission decision is made by each physician who is assigned a number of dedicated inpatient beds. ...This paper examines a decentralized admission control system with partial capacity sharing in a hospital setting. The admission decision is made by each physician who is assigned a number of dedicated inpatient beds. A physician can “borrow” beds from other physicians if his dedicated beds are all occupied. We seek to understand the impact of the “borrowing cost” on physicians’ admission behavior. We find that (i) If the borrowing cost is low, a physician tends to admit lower-risk patients when either his or others’ capacity utilization is higher;(ii) If the borrowing cost is moderate, a physician tends to admit higher (lower)-risk patients when his (others’) capacity utilization is higher;and (iii) If the borrowing cost is high, a physician tends to admit higher-risk patients when either his or others’ capacity utilization is higher. We then empirically test and validate these findings. Our work demonstrates that when designing strategic admission control systems, it is important to quantify and perhaps then influence the magnitude of the borrowing cost to induce a proper level of competition without sacrificing the benefit of resource pooling.展开更多
Distributed computing systems have been widely used as the amount of data grows exponentially in the era of information explosion. Job completion time (JCT) is a major metric for assessing their effectiveness. How to ...Distributed computing systems have been widely used as the amount of data grows exponentially in the era of information explosion. Job completion time (JCT) is a major metric for assessing their effectiveness. How to reduce the JCT for these systems through reasonable scheduling has become a hot issue in both industry and academia. Data skew is a common phenomenon that can compromise the performance of such distributed computing systems. This paper proposes SMART, which can effectively reduce the JCT through handling the data skew during the reducing phase. SMART predicts the size of reduce tasks based on part of the completed map tasks and then enforces largest-first scheduling in the reducing phase according to the predicted reduce task size. SMART makes minimal modifications to the original Hadoop with only 20 additional lines of code and is readily deployable. The robustness and the effectiveness of SMART have been evaluated with a real-world cluster against a large number of datasets. Experiments show that SMART reduces JCT by up to 6.47%, 9.26%, and 13.66% for Terasort, WordCount and InvertedIndex respectively with the Purdue MapReduce benchmarks suite (PUMA) dataset.展开更多
基金This work was supported by the National Natural Science Foundation of China(Nos.71720107003,72033003 and 71722008)。
文摘This paper examines a decentralized admission control system with partial capacity sharing in a hospital setting. The admission decision is made by each physician who is assigned a number of dedicated inpatient beds. A physician can “borrow” beds from other physicians if his dedicated beds are all occupied. We seek to understand the impact of the “borrowing cost” on physicians’ admission behavior. We find that (i) If the borrowing cost is low, a physician tends to admit lower-risk patients when either his or others’ capacity utilization is higher;(ii) If the borrowing cost is moderate, a physician tends to admit higher (lower)-risk patients when his (others’) capacity utilization is higher;and (iii) If the borrowing cost is high, a physician tends to admit higher-risk patients when either his or others’ capacity utilization is higher. We then empirically test and validate these findings. Our work demonstrates that when designing strategic admission control systems, it is important to quantify and perhaps then influence the magnitude of the borrowing cost to induce a proper level of competition without sacrificing the benefit of resource pooling.
基金This work was supported by the National Key Research and Development Project of China under Grant No.2020YFB1707600the National Natural Science Foundation of China under Grant Nos.62072228,61972222 and 92067206the Fundamental Research Funds for the Central Universities of China,the Collaborative Innovation Center of Novel Software Technology and Industrialization,and the Jiangsu Innovation and Entrepreneurship(Shuangchuang)Program.
文摘Distributed computing systems have been widely used as the amount of data grows exponentially in the era of information explosion. Job completion time (JCT) is a major metric for assessing their effectiveness. How to reduce the JCT for these systems through reasonable scheduling has become a hot issue in both industry and academia. Data skew is a common phenomenon that can compromise the performance of such distributed computing systems. This paper proposes SMART, which can effectively reduce the JCT through handling the data skew during the reducing phase. SMART predicts the size of reduce tasks based on part of the completed map tasks and then enforces largest-first scheduling in the reducing phase according to the predicted reduce task size. SMART makes minimal modifications to the original Hadoop with only 20 additional lines of code and is readily deployable. The robustness and the effectiveness of SMART have been evaluated with a real-world cluster against a large number of datasets. Experiments show that SMART reduces JCT by up to 6.47%, 9.26%, and 13.66% for Terasort, WordCount and InvertedIndex respectively with the Purdue MapReduce benchmarks suite (PUMA) dataset.
