Faculty Column Supercomputer: the solution to NP-hard problems? Optimization problems exist everywhere in real applications, including operating room scheduling in hospitals, vehicle routing in transportation, and protein folding prediction in computational biology. Many of them are NPhard, i.e. the computational time increases exponentially with the size of problem. You may think, as the computational power of computers developed so fast, we could be able to solve many problems that were impossible to solve in a reasonable amount of time. However, the evolution of modern sciences and technologies also lead to the demand of solution methods for large-scale systems, such as those in computational biology and in transportation systems. As you may know, the “deep blue” (from IBM) and more advanced computing machines have been applied in major airline companies. However, we still may have unpleasant experiences in our trips due to the bad operation planning and scheduling. Clearly, the supercomputer is not an all-cure solution and a powerful algorithm is more cost-effective than any supercomputer. Different from computer scientists, operations researchers study and develop many advanced algorithms for NP-hard optimization problems with solid mathematical analysis. The most famous is the branch-and-cut algorithm, which combines the classical branch-and-bound scheme and cutting planes to reduce the solution space in the solution procedure for mixed integer programs. In the past 15 year, the use of cutting planes theoretically and computationally has been fruitful. Cutting planes have proven to be very effective in deriving the optimal solutions within reasonable times. Many challenging real problems have successfully been solved using commercial solvers based on branch-and-cut algorithms, such as CPLEX and Xpress. These solvers are becoming the industrial standards—large oil companies, airline companies and a majority of the Fortune 500 have implemented CPLEX in their decision support systems. As a result, there exist many opportunities for operations researchers/industrial engineers with good skills on CPLEX. Currently, more and more attention has been put on complicated stochastic and fast real time/online optimization problems, such as those arising from airport control and service system scheduling. Some interesting questions naturally follow: How can we develop fast algorithms for these applications? Can we extend the wonderful work on cutting planes to them? We, operations researchers, will provide the answers ■
Alumni Q&A Session The following is a Q&A session was held with Mr. Abhik Bhattacharya, an alumnus of USF-IMSE (Graduated in May 2009). Could you tell us which company you currently work for, your position and briefly what your responsibilities are?
I recently joined Massachusetts General Hospital (part of the Partners Health System) in Boston, MA as an electronic inpatient record system (EMR) workflow analyst. My first/current project is geared towards helping to ensure that the planned partners (who include MGH, BWH & other hospitals) transition entirely to an EMR is at least time-neutral from the existing system (combination of paper & electronic) and identify workflow changes/issues arising from the change. What were the motivating factors for you to join the health care industry?
I always wanted to be in the service industry since I was a kid, and health care came as a natural progression. Let's just say I like assisting people to help other folks...and making the formers’ lives easy… What are the courses that you have found most applicable in your career so far?
Quite a few: Analysis of Health Care Issues (it gave a really great overview of the US health care industry), Advanced TQM (I believe that course is very important for someone pursuing this career path), Simulation (DES is incredibly useful along with a wow factor for non-technical audiences), Industrial Information Systems (for its great overview of a mixed bag of tools), SAS,... What advise can you give to the current IMSE graduate students?
USF-IMSE is a great department, with some of the most wonderful people. There are some great things going on that really help you in your career. I've known & met people from other schools and was amazed at how much more professional exposure & career development was available at USF-IMSE. So, you are at the right place, at the right time… However, I don't think I'm qualified to advise other people, since I'm still in the learning phase of my career, but I'll share what has worked for me so far (it might very well be the worst attitude to have!) - Do your own thing (be impervious to pressure). Make sure you enjoy what you do - so it never seems like work, and most importantly, never forget to have fun... Live your life your way! This interview was carried out by Wilkistar Otieno (OR Times Editor) via email. INFORMS USF would like to thank Mr. Abhik Bhattacharya, for his contribution to Volume 4, the OR Times.
Fall 2009
Issue 2
s issue: ….2 Inside thi ………………………… Service………
y of F Communit Management INFORMS@US Inventory me Ti l nia -i rt th Pa screte Product wi Optimal Di 3 riorating …………….………. te …… De …… le …… ng …… …… a Si …… …… …… …… ….3 gging ………… ………………………… By Dr. Bo Zeng Backlo e social…… om lc we E …….4 S and II ………………………… IMSE Assistant Professor INFORM al Meeting… nu An S RM FO IMSE at IN
A Newsletter from Student Chapter of INFORMS @ IMSE, USF
Community Service INFORMS@USF and Metropolitan Ministries of Tampa: IEs at your service! On late February of 2009, our INFORMS chapter received an invitation from Metropolitan Ministries (MM) to help them review their layout and operations flow for their 25,000 sq. ft. holiday tent. Every year in November and December, MM hosts a tent for distributing food and toys to those in need in Tampa (24,000 people served in 2008). Due to the ecoHoliday tent under construction. Taken from www.metromin.org nomic situation around the country, they expect a significant increase in solicitants and perhaps a decrease in donations for the 2009 holiday season. Consequently, there is only one solution: to become more efficient. After our first meeting, the indications were clear: people crowding at some specific points, not enough storage space, long distances walked by volunteers carrying diverse heavy items, and some activities that seemed not to add value to the process, but were needed for various reasons. We analyzed their current layout and detected a lot of opportunities to improve their distribution and simplify their process. After using traditional layout-distribution heuristics and discussing our with them, we were able to improve their layout. In the new layout, new queue areas were included, more flexible storage space was added, distances between clientcustomer areas were reduced, a single flow for the customer was designed, less volunteer workers are needed to deal with more clients, and safety for children and supplies were enhanced. You may want to stop by and see how good it looks now, so please go to Hillsborough Avenue in downtown Tampa to volunteer or donate this year, helping MM give others a moment of hope and joy. ■ By: Alfredo Santana-Reynoso, IMSE PhD. student
INFORMS@USF and the YEA After School Program: There is always a way to give! During the summer of 2008, a group of INFORMS colleagues and I were thinking on different ways to give back to the community. We decided that the best alternative was to offer tutoring and mentoring services in settings with limited assistance. We commented the initiative to Wilkistar O. who, due to her attitude towards service and involvement with several community initiatives, cheerfully embraced the idea and proposed the Youth Empowered to Achieve (YEA) program, which offers after-school services to low income students from underperforming elementary schools in the county. Since August 2008 to date, a group of INFORMS members including Chaitra G., Vishnu N., Patricio R., Laila C., Wilkistar O., Diana P. and Anahita H. have been spending two hours every week in helping the K-5 graders accomplish their mathematics, reading and writing homework. In the summer of 2009, INFORMS supported five of the students to participate in the STARS science summer camp—a one week camp organized by the Industrial Engineering department using an NSF grant. In this event, children were able to learn and recreate some of the most famous inventions of all-time. Also, in July 2009, the YEA program had a month of multicultural emphasis, where INFORMS members presented different aspects of their countries and cultures. We made presentations about the countries of Colombia (Laila C. & Diana. P.), India (Chaitra G. & Vishnu N.), Turkey (Fethullah C. & Ozan O.), Puerto Rico (Dayna M.), Kenya (Wilkistar O.) and the culture of Cherokee Indians (Courtney F.). This academic year, INFORMS@USF plans to continue and enlarge the support to the YEA program by encouraging new members to participate. The YEA program is a unique opportunity to share and enjoy with the children and to assure that, even with a small contribution, young citizens are encouraged to enjoy learning, develop critical thinking and become more tolerant to the diversity in their community. Written by: Diana Prieto. Graphic material: Wilkistar Otieno. IMSE Ph D. students
2
Research Corner Optimal Discrete-in-Time Inventory Management of a Single Deteriorating Product with Partial Backlogging The implicit assumption in conventional inventory models is that the stored products maintain the same utility forever, i.e., they can be stored for an infinite period of time without losing their value or characteristics. However, generally speaking, almost all products experience some sort of deterioration over time. Some products have very small deterioration rates, and thus the effect of such deterioration can be neglected. Some products may be subject to significant rates of deterioration. Fruits, vegetables, pharmaceuticals, alcohol, and radioactive materials are examples that can experience significant deterioration during storage. Therefore the effect of deterioration must be explicitly taken into account when developing inventory models for such products. In general, deterioration is defined as decay, damage, spoilage, evaporation, obsolesce or loss of utility of an item such that it cannot be used for its original purpose. The inventory models dealing with deterioration can be classified into two categories. The first category includes models with age-dependent ongoing deterioration (i.e., the items have fixed life-times). Such models are referred to as perishable inventory models. The second category consists of models with age-independent ongoing deterioration (i.e., the products have random life-times). Such models are called deteriorating inventory models. Milk, fish and blood are examples of category one, while perfume, alcohol and gasoline are examples of category two. It has been pointed out that due to spoilage the $1.7 billion apple industry in the U.S. loses as much as $300 million every year and the top 40 retailers in the U.S. dump as much as 500 million pounds of food every year (Andrew Webb, 2006). Therefore, an essential analysis of the effect of deterioration for inventory management of products that deteriorate is required. The objectives of this research are two-fold: (i) to identify the optimal ordering policy structures for a single deteriorating product over a discrete, finite/ infinite horizon under various conditions such as different demand patterns, time-varying deterioration rates and waiting-timedependent partial backlogging ratios, with minimum expected overall ordering, holding, deteriorating, backlogging and penalty costs; and (ii) to build mixed-integer programming models and develop efficient computational algorithms to solve more general cases, with fixed ordering costs, non-homogeneous demands, and other characteristics. Through computational experiments and sensitivity analysis, a thorough and insightful understanding of deteriorating inventory management will be achieved. ■ By Yang Tan, IMSE PhD. student http://communication.howstuffworks.com/how-inventory-management-systems-work.htm
3
IMSE at INFORMS 2009, San Diego Here is a summary of our presentations at the upcoming INFORMS annual meeting at San Diego, CA. Sunday Oct 11, 11:00 - 12:30 Tapas Das, Patricio Rocha, Ehsan Salimi A Two-tier Matrix Game Approach for Obtaining Joint Bidding Strategies in FTR and Energy Markets Wilkistar Otieno, Geoffrey Okogbaa
Monday Oct 12, 08:00 - 09:30
Reliability Modeling of Ge/ High-k Based MOS devices Patricio Rocha, Tapas Das, Vishnu Nanduri
Monday Oct 12, 11:00 - 12:30
Generation Capacity Expansion in Restructured Power Markets under a CO2 Cap-and-Trade Program Monday 13:30 - 15:00 Alcides Santander Mercado, Jose Zayas-Castro Identification of Patient Recovery Patterns after Cardiac Surgery based on Laboratory Tests Results Monday 13:30 - 15:00
Diana Prieto, Tapas Das
Volume 4, Issue 2
Real-time Applicability of Pandemic Modeling Approaches Qingwei Li, Alex Savachkin
Tuesday Oct 13, 13:30 - 15:00
The p-Median Location Problem: A Counter-measure Policy to Mitigate Random Facility Disruptions Athina Brintaki, Susana Lai-Yuen
Interactive Session Tuesday, Oct. 13 2009, 12:30 - 01:30 pm
Poster: “BioDE: A Novel Differential Evolution Approach for Molecular Conformational Search” Andres Uribe-Sanchez, Alex Savachkin
Tuesday Oct 13, 13:30 - 15:00
Developing Federal Resource Allocation Strategies to Mitigate Cross-regional Pandemic Outbreaks Dayna Martinez, Tapas Das Impact of Social and Behavioral Issues on Pandemic Influenza Containment
Wednesday Oct 14, 12:45 - 14:15
Anahita Hassanzadeh, Bo Zeng Dynamic Ambulance Relocation Model
Wednesday Oct 14, 14:45 - 16:15
Laila Cure, Jose Zayas-Castro
Wednesday Oct 14, 14:45 - 16:15
A Model for the Evaluation of Patient Safety Interventions
Fall 2009
Andres Uribe-Sanchez, Alex Savachkin
Wednesday Oct 14, 16:30 - 18:00
Analysis of Stochastic Disruptions to Support: Design of Capacitated Engineered Networks
Upcoming Events IMSE and INFORMS @ USF lecture series will welcome the following guest speakers: Dr Joseph Hartman: - University of Florida http://www.ise.ufl.edu/hartman/
Dr. Amy Cohn - University of Michigan http://www-personal.umich.edu/~amycohn/
Monday, October 5th 2009
TBA
Check out their research and come ready to learn and ask questions !
IMSE 4202 E. Fowler Ave. ENB 118 Tampa FL, 33620 Tel: (813) 974-5591 Fax: (813) 974-974-5953
[email protected]
Editors: • Meredith Bounds • Courtney Feliciani • • Wilkistar Otieno • Laila Cure• http://informs.eng.usf.edu 4