PITA Fiscal Year 2007 Projects

Product and Process Design and Optimization

Computational Models and Algorithms for Enterprise-wide Optimization of Process Industries
The process industry is a key industrial sector in the U.S. and in the state of Pennsylvania. For instance, the chemical industry is the major producer in the world (25% of world production) with shipments reaching $506 billion and a record $109.3 billion in exports in 2004. Enterprise-wide optimization (EWO) has become a major goal in this industry due to the increasing pressure for remaining competitive in the global marketplace. EWO involves simultaneously optimizing the operations of supply, manufacturing and distribution activities of a company to reduce costs and inventories. A major focus in EWO is the scheduling of manufacturing facilities, as well as their modeling at the proper level of detail, often requiring nonlinear process models. Major operational items include planning, scheduling, real-time optimization and inventory control. One of the key features in EWO is integration of the information and decision-making among the various functions that comprise the supply chain of the company. This is being achieved with modern IT tools, which together with the Internet has promoted e-commerce. To fully realize the potential of transactional IT tools, the development of sophisticated deterministic and stochastic linear/nonlinear optimization models and algorithms (analytical IT tools) is needed to explore and analyze alternatives of the supply chain to yield overall optimum economic performance, as well as high level of customer satisfaction. An additional challenge is the integrated and coordinated decision-making across the various functions in a company (purchasing, manufacturing, distribution, sales), across various geographically distributed organizations (vendors, facilities and markets), and across various levels of decision-making (strategic, tactical and operational).

It is the goal of this proposal to continue in the third year (July 1, 2007-June 30, 2008) the research work for developing a comprehensive set of computational capabilities for EWO problems. The work is performed by a multidisciplinary team from three institutions of Pennsylvania (Carnegie Mellon, Lehigh, U. Pittsburgh), composed of chemical and industrial engineers, and operations researchers, (5 faculty, 8 Ph.D. students), working in coordination with ICES at Carnegie Mellon and ATLSS at Lehigh. Our group is developing novel models, algorithms, decomposition methods, and computational techniques in order to provide a new generation of analytical IT tools. The project involves close collaborations with ABB, Air Products and Chemicals, BP, Dow Chemical, and ExxonMobil, who are members of the Center for Advanced Process Decision-making (CAPD) at Carnegie Mellon. NOVA Chemicals has recently joined the group as well. We have identified case studies with each of these companies. The work involves scheduling of cranes in steel manufacturing (ABB), rescheduling of bulk gas production and distribution (Air Products), stochastic optimization of multiperiod production planning models (BP), simultaneous strategic and tactical planning in distributed batch plants (Dow), global optimization of multiperiod refinery models (ExxonMobil), and modeling and evaluation platform for a specialty polymer product (NOVA). We have meetings twice per year, and have organized a popular seminar series that is broadcast to the companies. Another important event is the INFORMS Annual Meeting that will take place in Pittsburgh on November 5-8, 2006, in which we have organized four sessions on Enterprise-wide Optimization. This will provide high visibility for this project and allow us to disseminate our research results at a national level. Our research results are available in the webpage http://egon.cheme.cmu.edu/ewocp/, that we have developed for this project. We describe plans of our work that include efforts to expand industrial internships and recruit additional companies that are based in Pennsylvania.