Hearin Center for Enterprise Science

 

Awards

Home

Accomplishments of

 

Members of the Center have made distinguished theoretical and applied research contributions to science, government, business and engineering. These contributions include the development, implementation, testing and commercialization of specialized solution methods for many classes of practical problems. This research has involved the development of new mathematical algorithms, computer science data structures, computer implementation techniques, and computational testing techniques.  These algorithms are currently being utilized by over 100 government agencies and companies, and their   This utilization has been credited with saving over 300 million dollars.

Principal research areas include applications of computers to the fields of optimization, decision support, industrial planning, financial analysis, systems design, multicriteria analysis, applied artificial intelligence, energy, natural resources planning, logistics, transportation and distribution, manpower planning, large scale allocation models.

COMPUTER IMPLEMENTATION ACCOMPLISHMENTS OF HEARIN CENTER MEMBERS

  • An integrated production, distribution and inventory planning system based on algorithmic advances in solving large scale embedded network problems. This research received an International Achievement of The Institute of Management Sciences, and companies have used these procedures with savings of over thirty million dollars. The original article reporting this work has now been reprinted in seven different volumes on systems design, and the modeling component of this work is being taught at universities around the country, including Columbia, Princeton, Wharton, and the University of Chicago.
  • A microcomputer system for scheduling operations and personnel that has successfully handled problems three orders of magnitude larger than any of its class by a decade of prior research. This research effort is widely cited as a pioneering demonstration of the importance of the artificial intelligence/operations research interface and received the Best Application of Decision Science Theory Award of the Institute of Decision and Information Sciences.
  • Formulations and approaches for discrimination analysis that provide new methods for pattern recognition problems in artificial intelligence. This work has been implemented and tested by other researchers across the country and has been shown capable of solving discrimination problems that classical models cannot encompass, while yielding superior discrimination power over the classical models in their own realm of application. Recent developments are now using these approaches to train neural networks.
  • Expert analysis and computer solution procedures for production planning and distribution systems for GM Research Laboratories. This work provided fundamental breakthroughs in machine scheduling that have been incorporated into manufacturing of plastic molded parts throughout the industry.
  • An interactive microcomputer and graphics system for space planning and facilities layout design employing new techniques for AI structural analysis and pattern recognition. This system produced more than a hundredfold improvement in efficiency over previous procedures for these problems and is being routinely used by space planning companies such as Dalton, Dalton, Newport and Marshall Erdwin around the country. This work was also selected for citation in a survey of outstanding applications of microcomputer graphics systems.
  • An expert planning system to determine optimal lot-sizing and machine loading for multiple products used in multi-level planning of manufacturing operations. This work was implemented for a major U.S. manufacturing company and reported in a collection of published articles in Interfaces and AIEE Transactions on improved computer-based planning systems.
  • A modeling and computer solution system for determining optimal mining and ore extraction sequences for W. R. Grace, Inc.  The model encompasses an expert system component incorporating zero-one optimization to make decisions concerning depth and location of mining activities and has been implemented in the field since 1983.
  • A large-scale model and solution system for introducing new products and determining product distribution in the oil industry. This work solved large scale nonlinear and mixed integer programming problems that were previously unsolvable and appeared as the lead article in a volume dedicated to computer methods for industrial applications.
  • AI and OR models for plant layout design for General Motors Research Laboratories.  This work determined the subdivision of plant space and relative locations of assembly lines to inventories to enable efficient fabrication of final products.  GM credited this research with saving thousands of hours in engineering design, and with developing improved plant layout procedures that reduced both operating expense and capital costs of building.
  • A large scale expert system for allocating resources to meet demands of national and international emergency.  This system made it possible to make coordinated responses to conditions of national emergency in real time and improved the emergency response times by a factor of 600 to 1.

ENERGY AND RESOURCES PLANNING ACCOMPLISHMENTS OF HEARIN CENTER MEMBERS

  • A system for analyzing trade-offs among alternative energy sources and uses, with particular focus on exchanges between petrochemical and biomass based fuels, for the Solar Energy Research Institute.  The model and solution procedure of this system won an award from the Energy Research Institute for the analysis of energy issues, and has subsequently been expanded to a large scale national model published as a special invited paper in Energy Models and Studies.
  • A system for the U.S. Department of Agriculture for large scale forest planning operations over an eighty year planning horizon.  The system integrates the determination of economically feasible investment levels with the determination of policies for harvesting, transporting, clearing and re-planting of different types of timber to assure adequate supply and reserves for future needs.
  • A model formulation and solution methodology for scheduling nuclear refueling operations to coordinate use of electrical energy with hydroelectric and chemical energy sources for the Tennessee Valley Authority.  This work succeeded in generating schedules that improved on the schedules found by the best previous methods by over ten million dollars.
  • A procedure for optimally locating and sizing electrical power substations for the U.S. Department of Energy (ERDA), determining the most effective way to expand and contract electrical power facilities to meet the changing energy demands resulting from growth and population shifts.
  • A procedure for scheduling and coordinating the allocation of water at dams, reservoirs, and channels to maintain optimal levels and flows for hydroelectric and agricultural needs.  This procedure has been implemented in the United States, India, Sweden, the German Democratic Republic and Poland.

TRANSPORTATION AND DISTRIBUTION ACCOMPLISHMENTS OF HEARIN CENTER MEMBERS

  • A prototype of a Decision Support System (DSS) for a real-world tank truck dispatching problem which occur in a case study concerning the transportation of raw material throughout Europe. It was a consulting project for the Matra Cap Systems (France). This work resulted in a system that rapidly provided feasible an high quality solutions which were not possible to be obtained by general purpose optimization software.
  • Part of a Decision Support System (DSS) for the crew scheduling management of the Lisbon Subway. This is a complex crew scheduling problem (CSP) involving a number of additional constraints which do not occur in standard CSPs. This system reduced the number of duties for all the train timetables provided by the company.

NETWORK OPTIMIZATION ACCOMPLISHMENTS OF HEARIN CENTER MEMBERS

  • Partnered in creating the first efficient methods for solving networks with tens of thousands of equations and millions of variables using external storage devices.  This development has allowed the U.S. Military to substantially improve its human resource planning and assignment activities.  It has also allowed the U.S. Treasury to obtain merged micro-data files, which are substantially better for evaluating the fiscal impact of taxation, welfare, social security, etc., policy changes.
  • Efficient algorithms for solving multicriteria network flow problems. In 1975, the Assistant Secretary of Defense mandated that in order to reduce personnel turbulence "the Services will utilize simultaneous rather than sequential assignment selections in programming personnel reassignments."  Following this, the U.S. Army conducted an extensive study that credits these algorithms with making it possible to meet the mandate.  Their study showed that these algorithms made it possible to solve optimally multicriteria manpower assignment problems with 8,000 people and over a million eligible job assignments in less than 15 minutes. 
  • A generalized network model that has made it possible for the U.S. Government to develop a nationwide natural gas distribution model for evaluating national regulatory policies.  The Congressional records indicate that this code was 50 times more efficient that any other code for these problems.

HUMAN RESOURCES AND NATIONAL PLANNING ACCOMPLISHMENTS OF HEARIN CENTER MEMBERS

  • Personnel planning models and solution approaches for human resource planning and career path planning.  Due to the success of these approaches, the military services in 1982 issued a request for proposals totaling over 300 man years of effort which required the use of these models.
  • Modeling and solution of a problem for the U.S. Department of the Treasury and the National Bureau of Standards to determine categories, levels, and eligibility status for welfare payments, social security insurance payments and tax assessments.  Economic and social implications are analyzed by a model structure involving more than 62 million variables,  an order of magnitude larger than any comparable model previously considered by either government or industry.
  • Models and solution models for the Urban Mass Transit Authority of the Department of Transportation. The solution methods have proved capable of handling problems with thousands of junctions and hundreds of thousands of interconnections and demonstrated greater efficiency for the UMTA/DOT problems than any methods produced by two decades of prior research.
  • A study concerning the staff dimension and time tabling for a petrol port (SOPONATA) in Portugal. This study reduced the staff dimension while significantly improved the workforce (rostering) plan used by the company

Accomplishments