Universidad de Chile
Andrés Weintraub Pohorille is professor at the Department of Industrial Engineering of the School of Engineering, Universidad de Chile and Director of the Institute for Complex Systems in Engineering (ISCI).
He holds a degree in Electrical Engineering from the University of Chile and a Masters in Statistics and a Ph.D. in Industrial Engineering and Operations Research from the University of California, Berkeley. He is a Professor at the Department of Industrial Engineering, University of Chile. His main research areas are Operations Research and Operations Management in forestry and mining, logistics and transportation.
He has published over 70 papers in recognized journals, including Operations Research, Management Science, Forest Science, the European Journal of Operations Research. He has also edited several books and journal issues on topics related to use of Operations Research in natural resources. He has carried out multiple projects with industry and governmental organizations, including the US Forest Service and forest firms in Chile in models related to long range planning, short term harvesting and transportation, CODELCO, one of the largest copper firms in the world in models related to long range copper extraction, CSAV, a top 10 worldwide shipping company, housed in Chile, to improve the management of their 500.000 container business, the Chilean Ministry of Education, on determining best locations of rural schools, and the Chilean Football Association in scheduling the football season since 2005. The work with Chilean forest firms won the Edelman Prize, the most prestigious award for applied Operations Research, awarded by INFORMS, the US Society for Operations Research and Management Sciences. The work with CSAV was an Edelman finalist in 2011 and the football scheduling was a finalist in 2016.
He has led in the last 12 years the Institute for Complex Engineering Systems, which is currently funded yearly with 3 million dollars, and involves 60 researchers and a staff of 15 people. The Institute covers areas such as Operations Research, Data Science, Industrial Organization and Consumer Analytics.
The Institute is strong in developing projects with industry and the government which are original and have impact. Projects for which a financial impact could be measured reported increases in net revenues of 300 million dollars for 2010. The papers published by its researchers was ranked number one by Interfaces (the journal of INFORMS, the US Society of Operations Research and Management Sciences) in applied Operations Research publications.
He has received many recognitions which include: The Chilean National Prize for Applied Science in 2000 (which carries a lifetime monthly stipend of 1.200 dollars), the Harold Larnder Prize given by the Canadian OR Society, the INFORMS Presidential Prize, the Gold Medal from the Chilean Institute of Engineering, its highest recognition. He was awarded a Doctor Honoris Causa, from the University of Agricultural Sciences of Sweden, and the University of Laval, Canada. He is a member of the US National Academy of Engineering, and the Chilean National Academies of Science and of Engineering and an INFORMS Fellow.
He was a founder and former President of ALIO, the Latin American Association of OR, and President of IFORS, the International Federation of Operations Research Societies, which includes 50 country members, for the years 1998 to 2000.
Is it necessary to incorporate uncertainty into decision models?
All decision problems have a quota of uncertainty involved. These may be, among others, in terms of market demands, production rates, transportation times, catastrophes. One question a decision maker needs to answer is how to consider these uncertainties when developing decision making models. One option is to just disregard it and work with expected values. A more conservative simple option is to incorporate some ‘safety cushion’ , that is, to consider for example a production rate somewhat lower than the expected value. But ideally , uncertainty should be incorporated directly into the decision models. While there is an ample literature on methodological proposals to handle these problems, in practice there are multiple fields where such sophistication in the decision processes has not been developed (finance is one of the exceptions). In the last decade the developments in hardware and algorithms have allowed to consider improved techniques to develop stochastic models for real problems. The challenges in these developments are multiple. How to define the uncertainties in an exact way? How to formulate and solve these problems? How to determine if it is worthwhile to incorporate explicitly uncertainty given the high level of complexity it adds? In this talk we discuss these points in reference to concrete problems in logistics and natural resources.