Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SC47

3 - Acceptance Sampling for Surveillance of Forest Pests Robert G. Haight, US Forest Service, 1992 Folwell Ave, Saint Paul, MN, 55108, United States, Denys Yemshanov, Robert Venette, Cuicui Chen, Ning Liu, Frank H. Koch, Christian MacQuarrie, Krista Ryall Surveillance of forest pests helps foresters plan for pest control. We use acceptance sampling to delimit the extent of an infestation. The forester identifies potential survey sites and chooses the number of trees to inspect in each site given its tree density, infestation rate, and detection rate. If one or more trees in the sample is infested, the site is subject to pest control. The objective is to allocate a fixed survey capacity to sites to maximize the number of detected infested sites subject to an upper bound on the expected number of undetected infested trees. We formulate this objective as a MIP problem and apply it to surveillance of emerald ash borer, a destructive forest pest in Winnipeg, Canada. 4 - Modelling and Solving Biodiverstity Conservation Plans under Uncertainty Limited conservation budgets require prioritizing which management actions to implement and where to maximize the long-term persistence of biodiversity. Modelling this problem considering multiple species and actions is not an easy task. In addition, there is uncertainty in the data use for modelling and solving this problem (e.g. presence of the species, response to the action...). We propose a MIP-based framework for modeling and solving this problem. We seek for plans that maximize expected ecological benefit while minimizing spatial fragmentation and considering budget restrictions. n SC49 North Bldg 230 Supply Chain Planning in the Petrochemicals Sector Sponsored: Energy, Natural Res & the Environment/Natural Resources Petrochemicals Sponsored Session Chair: Ethan Malinowski, SUNY-Buffalo, Buffalo, NY, 14216, United States 1 - Dynamic Programming Approach for Maintenance Operations Budgeting Alcides Santander-Mercado, Associate Professor, Universidad del Norte, km 5 v a Puerto Colombia, Barraquilla, Colombia, Catalina Torres-Diaz Budgeting in industrial maintenance operations has been approached using forecasting methods and machine live cycle techniques. However, it is commonly ignored how the cost change overtime due technology upgrades, labor or availability of replacement parts. The comprehensive analysis of the maintenance cost and the new technology offered by their suppliers will help managers to decide whether continue with the original maintenance plan, draw by the manufacturer, or replace the machine. This research presents a Dynamic Programming approach to address the budgeting of maintenance operations, including a case study related to an application of this approach in a chemical company. 2 - Supply Chain Planning for Production Networks with Highly Flexible Modular Plants Tristan Becker, Ruhr University Bochum, Bochum, Germany, Stefan Lier, Brigitte Werners Modular production concepts are investigated to meet increased variability of product demands in process industries. Modular plants consist of standardized process modules, installed in containers, which allow for quick assembly, disassembly and relocation of production plants. Depending on the combination of process modules, different production capabilities are associated with a production plant. We apply new mixed integer programming formulations for production network planning under consideration of the new flexibility options to a real-world case study from the chemical industry. 3 - A Practical Approach to the Vehicle Routing Problem in Cylinder Gas Distribution Soma Toki, Tokyo Gas Co., Ltd., 1-5-20 Kaigan, Minato-ku, Tokyo, 105-8527, Japan, Naoshi Shiono, Eiji Murakami Recently, companies can remotely monitor the quantity of daily propane usage and plan their daily distribution operations with communication devices. In the liquefied petroleum gas (LPG) logistics industry, it is desirable that delivery plans for a week or more are always prepared. Companies can prospect a quantity of the jobs in the future with this plan. But when a LPG company has much of the distribute destinations, it is unrealistic to calculate the plan in one time. The company is required to disperse or reduce the calculation load for making the plans. In this work, we introduce some approaches to install the software for one LPG distributor. Jordi Garcia-Gonzalo, Forest Sciences Centre of Catalonia (CTFC), Solsona, 25280, Spain, Eduardo lvarez-Miranda, Andres P. Weintraub, Virgilio Hermoso, Jose Salgado-Rojas

n SC47 North Bldg 229A Behavioral Challenges in Policy Analysis with Conflicting Objectives Emerging Topic Session Chair: Margaret M. Wiecek, Clemson University, Mathematical Sciences Dept, Clemson, SC, 29634-1907, United States 1 - Behavioral Challenges in Policy Analysis with Conflicting Objectives Gilberto Montibeller, Loughborough University, School of Business & Economics, Loughborough, LE11 3TU, United Kingdom Public policy problems are rife with conflicting objectives: efficiency versus fairness, technical criteria versus political goals, costs versus multiple benefits. Multi-Criteria Decision Analysis provides robust methodologies to support policy makers in making tough choices and in designing better policy alternatives when considering these conflicting objectives. However, there are important behavioral challenges in developing these models. Policy analysis works with groups of policy makers, modeling their decision, facilitating their discussions, and representing preferences and priorities. The overarching goal is to improve decision processes and provide support to evidence-based decision making, taking into account public priorities and the inherent uncertainties that long term horizons and complex systems present. Key challenges in those interventions are the use of expert judgments, whenever evidence is not available, the elicitation of preference and priorities from policy makers and communities, and the effective management of group decision processes. Human behavior has a major influence on each of these challenges: experts might be biased in their estimates, individuals may be unable to express clearly their preferences, and groups may present dysfunctional dynamics. Extensive developments in behavioral decision research, social psychology, facilitated decision modeling, and incomplete preference models shed light on how decision analysts should address these issues to provide better decision support and develop high quality decision models. This tutorial discusses the main findings of these extensive, but rather fragmented, literatures. These guidelines are illustrated using policy analysis interventions conducted over the last decade for several organizations, such as the evaluation of capabilities of health systems against rabies for the World Health Organization (WHO), the prioritization of low moisture foods for the Food and Agriculture Organization of the United Nations (FAO), the assessment of bio-security threats for the UK Department of Environment Food and Rural Affairs (DEFRA), the evaluation of malaria treatment kits for the Malaria Consortium/USAID, and the prioritization of value-for-money studies for the UK National Audit Office. n SC48 North Bldg 229B Forest Management and Conservation Sponsored: Energy, Natural Res & the Environment Forestry Sponsored Session Chair: Jordi Garcia-Gonzalo, Forest Sciences Centre of Catalonia (CTFC), Ctra St Lloren de Morunys km2, Solsona, 25280, Spain 1 - PRISM - Harvest Scheduling using the 2012 Planning Rule The US Forest Service developed the Spectrum software in the mid 1990’s to help determine estimates of sustainable timber harvest volume as required by the 1976 National Forest Management Act . In the ensuing years the computing environment has changed. Additionally the 2012 Planning Rule updated the US Forest Service’s interpretation of the Act. The Northern Region in collaboration with Colorado State University developed a new linear programming model generator called PRISM to address these changes. PRISM uses a hybrid Model I and Model II formulation with a goal programming objective. The software uses an open source philosophy. Interface design and results from two Forest Plans are shown. 2 - Prioritizing Restoration of Fragmented Landscapes for Wildlife Protection: A Graph-theoretic Approach Denys Yemshanov, Natural Resources Canada, 1219 Queen Street East, Sault Ste. Marie, ON, P6A2E5, Canada, Robert G. Haight, Frank Koch, Mark-Andre Parisien, Tom Swystun, Barber Quinn, Cole Burton, Ning Liu, Salimur Choudhury Restoring habitat connectivity is critical for wildlife conservation in fragmented landscapes. We propose a network-based habitat restoration model with connected area requirements and a budget constraint. The MIP formulation shares some similarities with a budget-constrained Generalized Steiner Network problem. The objective is to determine restoration strategies that maximize the connected habitat capacity accessible to a wildlife species in a fragmented landscape. We apply the model to habitat restoration for woodland caribou in boreal forest in Cold Lake area, Alberta, Canada. David Anderson, USDA Forest Service, Albuquerque, NM, United States, Dung Nguyen, Eric Henderson, Yu Wei

78