Informs Annual Meeting 2017

TE63

INFORMS Houston – 2017

TE61

Prognostics and Health Management (PHM) information in CBM policy. In this paper, simulation-based optimization is used to assess the benefits of PHM- enabled CBM policy under resource dependence. Also, a framework is developed, numerically illustrated, and compared with classic CBM, and traditional maintenance policies using ARENA-based discrete event simulation model. 2 - Resilience of Core-periphery Networks Babak Heydari, Assistant Professor, Stevens Institute of Technology, 1 Castle Point, Babbio Center, Hoboken, NJ, 07030, United States, babak.heydari@stevens.edu We discuss conditions for resilience of core-periphery networks, an important class of complex systems. A successful model that captures resilience needs to model the extent of performance loss due to disruption, the post-recovery performance level, and the relative speed of recovery. We present a network model that incorporate connection costs and direct and indirect benefits of connection, assumes efficiency at both pre-disruption and steady-state post disruption systems, and focuses on core-periphery structures. We investigate conditions under which core-periphery structures increase the overall resilience of the network. 3 - Optimal Selective Maintenance for Networked Infrastructure Systems Wanshan Li, Tsinghua University, Beijing, 100084, China, liws16@mails.tsinghua.edu.cn, Chi Zhang Networked infrastructures, such as electricity, telecommunication and transportation, are considered critical for both economic development and social wellbeing of modern societies. To ensure their continued effective performance, cost-effective maintenance without interrupting their operation is necessary. This research proposes a new selective maintenance approach that can help determine optimal maintenance strategies for networked critical infrastructures with general structures. Our approach can also ensure an infrastructure’s ability of satisfying the customers’ demand continuously, even during the process of maintenance. 4 - Economic Design of Repetitive Inspection Plan with Testing Errors Young H.Chun, Louisiana State University, 2211 Business Education Complex, Baton Rouge, LA, 70803-6316, United States, prof@drchun.net Inspection has been widely used as an important tool for quality improvement in manufacturing and service industries. However, inspection errors are inevitable during a screening process of manufactured items such as computer chips; some of the defective items are tested negative and other non-defective items are tested positive erroneously. To improve the outgoing quality, each item is often tested more than once. Based on the test results, we can estimate the defective rate and the inspector’s type I and II errors via the method of maximum likelihood. We also propose various types of repetitive inspection plans and compare their performances in a simulation study. 370D Simulation and Optimization Contributed Session Chair: Ashutosh Nayak, Purdue University, West Lafayette, IN, United States, nayak2@purdue.edu 1 - A Decomposition Method for the Two Stage Staffing Problem in Multiskills Call Centers under Arrival Rate Uncertainty Thuy Anh Ta, University of Montreal, CP 6128 We consider a stochastic staffing problem with uncertain arrival rates. The objective is to minimize the total cost of agents under some chance constraints, defined over the randomness of the service level in a given time period. In the first stage, an initial staffing must be determined in advance. At a later time, when the forecast becomes more accurate, this staffing can be corrected with recourse actions by adding or removing agents. We propose a decomposition algorithm based on simulation, cut generation and stochastic programming to solve the problem in reasonable computational time. 2 - Selecting Good Enough Designs using Optimal Computing Budget Allocation Fei Gao, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong, feigao9-c@my.cityu.edu.hk, Siyang Gao We study the problem of selecting a subset of good designs from a finite set of simulated designs. We develop an approach to select r good enough designs instead of the exact top r designs, where good enough designs are defined as the top g designs (r ≤ g). Using the optimal computing budget allocation framework, we formulate the problem as that of maximizing the probability of correctly selecting r good enough designs under a simulation budget constraint. Based on the approximate measure of the probability of correct selection, we derive an asymptotically optimal selection procedure for selecting good enough designs. The proposed method shows good performance on some typical selection problems. TE63 Succursale Centre-Ville, Montreal, QC, H3W1C5, Canada, tathuyanh1989@gmail.com, Wyean Chan, Fabian Bastin, Pierre l’Ecuyer

370B Finance, Portfolio Analysis Contributed Session Chair: Aloagbaye Momodu, University of Toronto, Toronto, ON, Canada, aimomodu@mie.utoronto.ca 1 - Portfolio Construction using Gerber Statistics Based Risk Parity Model Ravi Shankar, Indian Institute of Technology Delhi, New Delhi, India, ravi@dms.iitd.ac.in, Dhanya Jothimani, Surendra S. Yadav In this research, we use Hierarchical Risk Parity model for portfolio optimization, where covariance of stocks are clustered. Covariance of the stocks is obtained using Gerber statistics. Sample consisted of firms listed on National Stock Exchange during the period 2008-2016. We find that the performance of the proposed model was better than Global Minimum Variance model. The contributions are two-fold: first, the model helps to avoid errors arising from estimation of returns and second, it avoids the need for inversion of covariance matrix. 2 - Identifying Robust Diversified Portfolios with Second Order Stochastic Dominance Constraints Peng Xu, PhD Candidate, Aalto University School of Business, Runeberginkatu 22-24, Helsinki, FIN-00100, Finland, peng.xu@aalto.fi Identifying Second-order Stochastic Dominance (SSD)-efficient portfolios is of great interest to finance research where investors are routinely assumed to be rationally risk-averse in making financial decisions. This paper seeks to (i) evaluate some most recent SSD approaches and their out-of-sample performance in portfolio diversification, and (ii) examine if robust optimization improves the out-of-sample performance of these approaches. We report the results from an empirical application analyzing how robust optimization based diversification among industry portfolios increases the likelihood of obtaining out-of-sample SSD over the market portfolio. 3 - Solving Cardinality Constrained Mean – Variance Portfolio Optimization using Message Passing Algorithms Alexia Yeo, University of Toronto, Toronto, ON, Canada, ayeo@mie.utoronto.ca The cardinality constrained mean-variance optimization problem has the advantage of reducing transaction costs and the complexity of asset management. It may be see as a special case of the quadratic knapsack problem. We approach this problem by using probabilistic inference in a graphical model to estimate portfolio selections. Message passing is a class of inference algorithms that utilizes the factorization and decomposition properties of the problem’s graphical model to efficiently estimate the value of unknown parameters. We evaluate the performance of the message passing algorithms MPLP and TRW against branch- and-cut algorithms. Contributors: Alexia Yeo, Dr. Roy Kwon 4 - Multiasset Game Contingent Claims Valuation using a Projected Successive Over Relaxation Algorithm Aloagbaye Momodu, University of Toronto, 121, Wolseley Street, Toronto, ON, M6J.1K1, Canada, aimomodu@mie.utoronto.ca, Chi-Guhn Lee Game contingent claims (Israeli options) are generalizations of American contingent claims in which the writer (seller) has the right to cancel the contract before maturity subject to the payment of a prespecified penalty. This added feature leads to complexities in the valuation with no analytical solution in the literature for finite maturity game contingent claims. In this paper, we value multi-asset game contingent claims in a standard Black-Scholes framework by solving finite difference discretization of the Black-Scholes partial differential equation using a projected successive over relaxation algorithm. 370C Resource Allocation Contributed Session Chair: Young Chun, Louisiana State University, Baton Rouge, LA, United States, prof@drchun.net 1 - Simulation-based Optimization of Prognostics and Health Management Enabled Condition-based Maintenance Policy Taiwo Joel Omoleye, Mr, City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Kowloon, Hong Kong, jtomoleye3-c@my.cityu.edu.hk, Kwok Leung Tsui The practical implementation of condition-based maintenance (CBM) is trailing behind the widespread theoretical benefits in the literature. This necessitates the growing concern about the effectiveness of CBM policy in industries implementing CBM. To overcome this, research is focusing on the value of TE62

433