Under the support of COST, I have got a chance to go to Spain for a research visit from April 12th, 2011 to April 18th, 2011. During my visit to Spain, I have met Prof. M.A. Hinojosa from Universidad Pablo de Olavide in Sevilla, Prof. A.M. Conde from Universidad de Sevilla and Prof. D. Ruan from Belgian Nuclear Research Centre (SCK) during his visit to Universidad Complutense de Madrid. I have discussed several research issues with the professors, and the outcome of the discussion is summarized as follows:
The discussion on research issues in Sevilla with Prof. Hinojosa and Prof. Conde has 2 main categories: 1) Multi-Criteria Decision Analysis (MCDA) problems under partial information. According to the research conducted by the 2 professors, partial information in the MCDA problems is reflected by weights of each criterion, i.e., the weight is not a precise value, but a possible value within a feasible region modelled by a set of linear constraints, based on which, the MCDA problem is solved by an optimization model with the performance of the alternative as an objective and the constraints on weights as major constraints. 2) Combination of cross evaluation approach in Data Envelop Analysis (DEA) with game theory. To overcome the limitations of traditional CCR model in DEA regarding the existence of unreasonable weights, cross evaluation approach was proposed. However, such an approach also has its limitations, which were mainly caused by non-uniqueness of optimal weights. In addition, the fact that there are always direct/indirect competitions among Decision Making Units (DMUs) is not considered either. Therefore, the idea of game theory is introduced into DEA. By solving the classical Nash bargaining model, the solution can not only overcome the limitations mentioned above, it can also generate a Pareto Optimum to provide motivation to all DMUs to follow the strategy which canlead to the optimal solution.
On the other hand, my research is about MCDA under uncertainties with applications in risk/security analysis in different social and industrial settings, and the implications from discussions mentioned above regarding my research include: 1) the way to represent and handle partial information proposed by the professors in Sevilla is quite similar to the idea I proposed in Manchester regarding the way to handle incompleteness during the process of Container Line Supply Chain (CLSC) security evaluation, according to which, the incompleteness is reflected by linear constrains on activation weights of different belief rules. Therefore, more in-depth discussion is necessary and cooperation on this research issue is possible in the future. 2) The idea of combing game theory with DEA is also useful in my research: in security analysis in CLSC, especially in optimal resource allocation problem in which limited resources are allocated to different areas to improve security of a certain stakeholder along a CLSC, there are competitions among the areas whose security need to be improved; further, the allocation problem itself can be considered as a DEA model in which each area can be seen as a DMU with resources required as input and benefit generated due to the improvement of security as output. However, there are still problems for this idea, such as how to quantify the input and output in the DEA model or how to modify the DEA model to make it capable to handle different forms of information. Therefore, more discussions on this issue are needed in the future.
The discussion on research issues in Madrid with Prof. Ruan mainly focuses on current research conducted in SCK regarding safety/security analysis and potential future research directions of the topic, especially after the recent incident regarding nuclear plant caused by earthquake in Japan. The main research issues in safety/security analysis currently in SCK is how to conduct the analysis under incomplete information, and how to generate a warning message before the actual disaster based on available information. As I have done research in the area of risk/security analysis, especially risk/security analysis under uncertainties, for several years, the discussion is very helpful for me to know how security analysis is conducted in other research groups, and more importantly, what current problems are faced by researchers in security analysis, and what the future direction may be for risk/security analysis under uncertainties.
The discussion on research issues in Sevilla with Prof. Hinojosa and Prof. Conde has 2 main categories: 1) Multi-Criteria Decision Analysis (MCDA) problems under partial information. According to the research conducted by the 2 professors, partial information in the MCDA problems is reflected by weights of each criterion, i.e., the weight is not a precise value, but a possible value within a feasible region modelled by a set of linear constraints, based on which, the MCDA problem is solved by an optimization model with the performance of the alternative as an objective and the constraints on weights as major constraints. 2) Combination of cross evaluation approach in Data Envelop Analysis (DEA) with game theory. To overcome the limitations of traditional CCR model in DEA regarding the existence of unreasonable weights, cross evaluation approach was proposed. However, such an approach also has its limitations, which were mainly caused by non-uniqueness of optimal weights. In addition, the fact that there are always direct/indirect competitions among Decision Making Units (DMUs) is not considered either. Therefore, the idea of game theory is introduced into DEA. By solving the classical Nash bargaining model, the solution can not only overcome the limitations mentioned above, it can also generate a Pareto Optimum to provide motivation to all DMUs to follow the strategy which canlead to the optimal solution.
On the other hand, my research is about MCDA under uncertainties with applications in risk/security analysis in different social and industrial settings, and the implications from discussions mentioned above regarding my research include: 1) the way to represent and handle partial information proposed by the professors in Sevilla is quite similar to the idea I proposed in Manchester regarding the way to handle incompleteness during the process of Container Line Supply Chain (CLSC) security evaluation, according to which, the incompleteness is reflected by linear constrains on activation weights of different belief rules. Therefore, more in-depth discussion is necessary and cooperation on this research issue is possible in the future. 2) The idea of combing game theory with DEA is also useful in my research: in security analysis in CLSC, especially in optimal resource allocation problem in which limited resources are allocated to different areas to improve security of a certain stakeholder along a CLSC, there are competitions among the areas whose security need to be improved; further, the allocation problem itself can be considered as a DEA model in which each area can be seen as a DMU with resources required as input and benefit generated due to the improvement of security as output. However, there are still problems for this idea, such as how to quantify the input and output in the DEA model or how to modify the DEA model to make it capable to handle different forms of information. Therefore, more discussions on this issue are needed in the future.
The discussion on research issues in Madrid with Prof. Ruan mainly focuses on current research conducted in SCK regarding safety/security analysis and potential future research directions of the topic, especially after the recent incident regarding nuclear plant caused by earthquake in Japan. The main research issues in safety/security analysis currently in SCK is how to conduct the analysis under incomplete information, and how to generate a warning message before the actual disaster based on available information. As I have done research in the area of risk/security analysis, especially risk/security analysis under uncertainties, for several years, the discussion is very helpful for me to know how security analysis is conducted in other research groups, and more importantly, what current problems are faced by researchers in security analysis, and what the future direction may be for risk/security analysis under uncertainties.