Decision Space Problem Space vs Data Space Motivation

Decision Space & Problem Space vs Data Space

Motivation New understanding requires Synthesis Belief Revision

Decision Aiding has generic principles worthy of research

Decision Space Shapes Treatment of the Data Space Data Conditioning Clustered Explicit Relevance Implicit Relevance SNR improvement Interpreted with respect to decision criteria Transformation (Interpretation) Decision Space Rules Initial Conditions Biases

State unknown Nested Heterogeneous Subattributes Known clear Unconfirmed/Suspected Red Macro attributes of each object (cell) Known Red SP Subject to Red Fires Undetected Red Drilling down -- Qualitative Decomposition Summary Indicator Weapons Personnel Ammo Transport Morale Fuel Fuzzy Sets treatment enables greater qualitative resolution Fully Capable Almost Fully Capable Partially Capable Almost Incapable

Tempus Fugit Critical Open Question How soon must I decide? • Representation of time • Representation of risk of no decision • Feedback into transformation process Identify which decision inputs subject to change • When do I know enough about what matters? • When will I not know more about what matters? • Manage the benefit/risk of delayed decison

Developing the Decision Space B A C D E Size of dot represents error associated with interpretation/assessment of data. Risk of bad decision increases as a large dot approaches a thin fuzzy boundary. Risk is greatest when large dot is near boundary of multiple decision choices, as in triple points ACD, ABC, DCE, BDE Decision D, as drawn, represents an option contiguous to all the other options, but not much room for error in the choice.

Help Wanted It is not economically feasible to develop generic technologies when customers demand proof of principle against specific applications … And then don’t provide a suitable database!!! Solution: Produce standardized, multidomain, entrylevel problem