1: \begin{abstract}

2: We propose an computational framework for real-time risk assessment

3: and prioritizing for random outcomes without prior information on

4: probability distributions. The basic model is built based on satisficing

5: measure (SM) which yields a single index for risk comparison. Since

6: SM is a dual representation for a family of risk measures, we consider

7: problems constrained by general convex risk measures and specifically

8: by Conditional value-at-risk. Starting from offline optimization,

9: we apply sample average approximation technique and argue the convergence

10: rate and validation of optimal solutions. In online stochastic optimization

11: case, we develop primal-dual stochastic approximation algorithms respectively

12: for general risk constrained problems, and derive their regret bounds.

13: For both offline and online cases, we illustrate the relationship

14: between risk ranking accuracy with sample size (or iterations).\\

15: \emph{}\\

16: \emph{Keywords:} Risk measure; Satisficing measure; Online stochastic

17: optimization; Stochastic approximation; Sample average approximation;

18: Ranking; 

19: \end{abstract}

20: