On a preference relation between random variables related to an investment problem (Q6607780)

The paper under review is to study an optimal investment problem with huge difference between the deterministic case and stochastic case. The deterministic problem: assume that one monetary unit invested in a risk-free asset at moment \(s\) becomes \(\sigma (s, t)\ge 0\) at the moment \(t> s\) with \(\sigma (s, s) =1\). If an investor decides to invest \(x_i\) at time \(t_i\), then the value at future time \(t\) (where \(t_1< t_2 < t\)) is \(U_{12}(t) =x_1 \sigma (t_1, t) + x_2 \sigma (t_2, t)\); if the investor swaps \(x_i\), then the value would be \(U_{21} = x_2\sigma(t_1, t) + x_1 \sigma(t_2, t)\). Which one is more valuable? If \(x_1\le x_2\), then \(U_{21} \ge U_{12}\) provided \(\sigma (s, t)\) is nonnegative increasing in \(t\) and decreasing in s. This problem can be extended to \(n\) portfolios with \(x_1\le x_2 \le \cdots \le x_n\) and \(\pi\) permutation of \(\{1, 2, \cdots, n\}\) and \(\sigma(t_1, t) \ge \sigma(t_2, t) \cdots \ge \sigma (t_n, t)\) for \(t_1\le t_2\le \cdots \le t_n\). Then \(U_e \le U_{\pi} \le U_{\omega}\) for \(\omega = \binom{1,2,\ldots ,n}{n,n-1,\ldots,1}\) is the reverse permutation. Hence, the best strategy is to invest the money in decreasing order. What is the best strategy if the amount of money to invested at each moment is stochastic? There are mean-risk analysis and stochastic dominance used in the stochastic portfolio investment. The paper under review is to take the stochastic dominance to address this problem.\par Section 2 illustrates the stochastic order, Hazard rate order and likelihood ratio order (see \textit{M. Shaked} and \textit{J. G. Shanthikumar} [Stochastic orders and their applications. Boston, MA: Academic Press (1994; Zbl 0806.62009)]) for more detail on stochastic orders and their applications), Authors defined a bank relation by the fact that \(aX+bY \le_{st} aY+bX\) for all \(a\ge b \ge 1\) and independent nonnegative r.v.s \(X\) and \(Y\), specifically for the purpose of the problem. Then, Proposition 2.4 shows that \(U_e \le U_{\pi} \le U_{\omega}\) again is true for the bank relation \(X_1\le_{bank} X_2 \le_{bank} \cdots \le_{bank} X_n\). The bank relation is just straightforward extension for the deterministic case, Proposition 2.5 discusses the interrelation between the bank relation and the other three orders, and the proofs of Proposition 2.4 and Proposition 2.5 are given in the Appendix. \par Section 3 tries to characterize the bank relation under the permutation action, and a few examples are given. Section 4 concludes the best strategy for the independent random amounts of money to invest, the question left for possible dependent r.v.s \(X\) and \(Y\).