Can Info-Gaps “localness” flaw be fixed by means of a robustness analysis at a number of different estimates spread over the complete region of uncertainty?
Answer-67: No it cannot. Suppose that we adopt this prescription and we conduct Info-Gap’s robustness analysis k times using k different estimates, û(1),…,û(k). We would then end with k (potentially different) optimal decisions, d*(1),…,d*(k), to reckon with. How would we decide which decision to select? In other words, instead of resolving the “localness” of Info-Gap’s robustness analysis, this attempted “rescue operation” brings out more forcefully how deep rooted and intractable a failure it is. The failure here is intractable because the basic approach to severe uncertainty is flawed. It fails to recognize that the real difficulty posed by the true value being subject to severe uncertainty is that for each value of the parameter there is (potentially) a different optimal decision. So, modifying Info-Gap’s robustness analysis by means of the above prescription would require that the modified analysis be additionally backed up by a theory prescribing the selection of decisions un
Related Questions
- Can Info-Gaps "localness" flaw be fixed by means of a robustness analysis at a number of different estimates spread over the complete region of uncertainty?
- Don you miss the message and intent of the Info-Gap approach (eg. robustness of solutions and local nature of the analysis?
- What is the fundamental flaw in Info-Gaps generic robustness model?
