Missing data processing model for unconventional emergency decision-making under incomplete information.

When the AHP and ANP are applied to the "scenario-response" based unconventional emergency decision-making, there easily exist inconsistent data and missing data because of time constraint, incomplete decision information, limited experience and human cognitive capabilities. In this paper, a logarithmic mean induced bias matrix model is proposed to process the missing data. Specifically, the missing data is first filled with unknown variables to obtain quasi-complete matrix, then the logarithmic mean induced bias matrix model is used to build overdetermined equations or constrained nonlinear optimization problems to solve the unknown variables. An emergency simulation experiment is used to illustrate and validate the proposed model, and the simulation results show that it is effective and feasible. The proposed model is only based on the original matrix information and does not need to calculate the weight vector. The estimated solutions of unknown variables can keep the global consistency and help emergency decision makers make valid decisions. [ABSTRACT FROM AUTHOR]