�ݺ�ߣ

International Journal of Business Marketing and Management (IJBMM)
Volume 8 Issue 2 Mar-Apr 2023, P.P. 91-101
ISSN: 2456-4559
www.ijbmm.com
International Journal of Business Marketing and Management (IJBMM) Page 91
Robust Optimal Reinsurance and Investment Problem
with p-Thinning Dependent and Default risks
Yingqi Liang1
, Peibiao Zhao1
1
School of Mathematics and Statistics,Nanjing University of Science and Technology, Nanjing Jiangsu
210094,China
843186823@
Abstract: In this paper, we consider an AAI with two types of insurance business with p-thinning dependent
claims risk, diversify claims risk by purchasing proportional reinsurance, and invest in a stock with Heston
model price process, a risk-free bond, and a credit bond in the financial market with the objective of maximizing
the expectation of the terminal wealth index effect, and construct the wealth process of AAI as well as the the
model of robust optimal reinsurance-investment problem is obtained, using dynamic programming, the HJB
equation to obtain the pre-default and post-default reinsurance-investment strategies and the display expression
of the value function, respectively, and the sensitivity of the model parameters is analyzed through numerical
experiments to obtain a realistic economic interpretation. The model as well as the results in this paper are a
generalization and extension of the results of existing studies.
Key words: p-Thinning dependent risk, defaultable bonds, dynamic programming, HJB equation, robust
optimization.
I. Introduction
The sound operation and solid claiming ability of insurance companies are important to maintain social
stability and promote social development, and the purchase of reinsurance can well control the risk of insurance
companies. The current research on reinsurance in the actuarial field of insurance mainly focuses on the insurer's
objective function, investment type, claim process, and price process of assets. [1]-[3]studied the most available
reinsurance investment problems under different utility function. [4]-[7]investigate the optimal reinsurance
investment problem under different risky asset price process.
Most of the current literature considers only single-risk claims, while in reality the risks are usually
correlated with each other. The current literature on claim dependency risk is divided into two main types of
dependency, one is co-impact and the other is p-thinning dependency. P-thinning dependency means that when
one type of claim occurs, there is a certain probability p that another type of claim will occur, for example, a car
accident or fire will not only cause property damage, but also loss of life if the situation is serious. The problem
under diffusion approximation model, jump diffusion risk model is studied in[8][9]，[10] studied the type of
dependent risk for common shocks, and [11] considered the risk process under p-thinning dependence. Most of
the current studies consider only two assets for considering investments in credit bonds. [12]-[14] take into
account defaulted bonds among the types of investments for insurers. Since models with diffusion and jump
terms usually introduce uncertainty in the model for insurers, insurers usually want to seek a more robust model.
[15]-[17] study the AAI most reinsurance-investment problem.
Based on this, this paper studies the optimal proportional reinsurance-investment problem of AAI under the
Heston model by considering both sparse dependence and claims risk based on [11][14][17]. The paper is
organized as follows: a robust optimal reinsurance-investment problem model under p-thinning dependence and
default risk is developed in Section 2. The value function is divided into pre-default and post-default in Section
3, and the optimal reinsurance-investment strategies are solved for pre-default and post-default using dynamic
programming, optimal control theory, and the HJB equation, respectively. Parameter sensitivities are analyzed,
and economic explanations are given in Section 4. Section 5 concludes the paper.
II. Model Formulation
Suppose *Ω, F, *𝐹𝑡+t∈,0,T -, P+ is a complete probability space, the positive number T denotes the final
value moment, [0, T ] is a fixed time interval, 𝐹𝑡 denotes the information in the market up to time t, and 𝔽 ∶=
*𝐹𝑡+t∈,0,T -denotes the standard Brownian motion 𝑊0(𝑡), 𝑊1(𝑡), 𝑊2(𝑡), 𝑊3(𝑡). Poisson process N(t), and the
right-continuous P-complete information flow generated by the sequence of random variables *𝑋𝑖, 𝑖 ≥ 1+,
*𝑌𝑖, 𝑖 ≥ 1+. ℍ ∶= (ℋ𝑡)t⩾0 is the information flow generated by the violation process H(t), let 𝔾: = (𝒢𝑡)t⩾0, be
the information flow generated by 𝔽, ℍ the expanded information flow, i.e., 𝔾: = ℱ𝑡 ∨ ℋ𝑡. By definition, each
𝔽 -harness is also the 𝔾 -harness. The probability measure P is a realistic probability measure and Q is a risk-
neutral measure. In addition, it is assumed that all transactions in the financial market are continuous, and no

Robust Optimal Reinsurance and Investment Problem with p-Thinning Dependent and Default ris
taxes do not incur transaction costs and all property is infinitely divisible.
1.1 Surplus Process
Assuming that the insurer operates two different lines of business and that there is a sparse dependency between
these two lines of business, the surplus process is as follows:
R(t) = 𝑥0 + 𝑐𝑡 − (∑ 𝑋𝑖
𝑁(𝑡)
𝑖=1
+ ∑ 𝑌𝑖
𝑁𝑝(𝑡)
𝑖=1
) . #(1)
where{𝑋𝑖，𝑖 ≥ 1} is independently and identically distributed in 𝐹𝑋(∙),𝐸(𝑋) = 𝜇𝑋 > 0，𝐸(𝑋2
) = 𝜍𝑋
2
, as the
claim amount of the first class of business，{𝑌𝑖，𝑖 ≥ 1} is independently and identically distributed in𝐹𝑌(∙
),𝐸(𝑌) = 𝜇𝑌 > 0，𝐸(𝑌2
) = 𝜍𝑌
2
, as the claim amount of the second class of business. the claim amount of the
second type of business. N(t)denotes the conforming Poisson process with parameter c denotes the premium of
the insurance company by the expected value premium there are c = (1 + θ1)λ𝜇𝑋 + (1 + θ2)λp𝜇𝑌.θ1 > 0,θ2 >
0.
1.2 Proportional Reinsurance
Assuming that the insurer diversifies the claim risk by purchasing proportional reinsurance, and let 𝑞1(𝑡),𝑞2(𝑡)
be the insurer's retention ratio, the claim after the insurer purchases reinsurance is:
𝑞1(𝑡)𝑋𝑖 , 𝑞2(𝑡)𝑌𝑖 then the reinsurance fee is δ(𝑞1(𝑡), 𝑞2(𝑡)) = (1 + η1)(1 − 𝑞1(𝑡))λ𝜇𝑋 + (1 + η2)(1 −
𝑞2(𝑡))λp𝜇𝑌.According to Grandell(1991)[8]
,the claims process can be diffusely approximated as:
d ∑ Xi
N(t)
i=1
= λE(Xi)dt − γ1dWX(t), γ1 = √λE(XI
2
)
d ∑ Yi
Np(t)
i=1
= λE(Yi)dt − γ2dWY(t), γ2 = √λE(YI
2
)
The correlation coefficient of 𝑊𝑋(𝑡)𝑊𝑌(𝑡 )is ρ
̂ =
λp
𝛾1𝛾2
E(𝑋𝑖)E(𝑌𝑖).Then the wealth process of the insurer after
joining the reinsurance is:
dXq1,q2 = ,λμX(θ1 − η1 + η1𝑞1(𝑡)) + λpμY(θ2 − η2
+η2𝑞2(𝑡))𝑑𝑡 + √(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2dW0
1.3 Financial Market
Suppose the financial market consists of three assets: risk-free assets, stocks, and corporate bonds, and the price
processes of the three assets are as follows: The price process of risk-free bonds is given by：𝑑R(t) =
rR(t)dt.The stock price process S(t) obeys the Heston stochastic volatility model:
{
𝑑𝑆(𝑡) = 𝑆(𝑡),𝑟 + 𝛼𝐿(𝑡)𝑑𝑡 + √𝐿(𝑡)𝑑𝑊1(𝑡)-, 𝑆(0) = 𝑠0
𝑑𝐿(𝑡) = 𝑘(𝜔 − 𝐿(𝑡))𝑑𝑡 + 𝜍√𝐿(𝑡)𝑑𝑊2(𝑡), 𝐿(0) = 𝑙0
R is the risk-free rate,α, 𝑘, ς, are positive constant.𝐸,𝑊1𝑊2- = ρt, 2𝑘𝜔 ≥ ς2
Following Bielecki (2007)[18]
, the credit bond price process p(t, T1), under a realistic measure P, using an
approximate model to portray default risk is as follows:
dp(t, T1) = p(t−, T1),(r + (1 − H(t))δ(1 − Δ)dt
−(1 − H(t−))ζdMp
(t)-
Where Mp
(t) = H(t) − hQ
∫ Δ(1 − H(u))du
t
0
is aℊ − harness， δ = 𝑕𝑄
ζ is the credit spread.
1.4 Robust optimization problem
Assuming the insurer adopts a reinsurance investment strategyε(t) = (𝑞1(𝑡), 𝑞2(𝑡), π(t), β(𝑡))，𝑞1(𝑡), 𝑞2(𝑡)are
the reinsurance strategies adopted by the insurer at moment t in the first and second asset classes,
respectively.π(t) for the insurance company's investment in equities at time t.，β(𝑡)for the insurance company′s
investment in credit bonds at time t. Let 𝔸 denotes the set of all feasible strategies, then the dynamic process of
wealth of the insurance company at this moment 𝑋𝜀
(t) is:
d𝑋𝜀(𝑡) = 𝜋(𝑡)
𝑑𝑆(𝑡)
𝑆(𝑡)
+ β(𝑡)
𝑑𝑝(𝑡)
𝑝(𝑡)
+ (𝑋𝜀(𝑡) − 𝜋(𝑡) − β(𝑡))
𝑑𝐵(𝑡)
𝐵(𝑡)

+𝑑𝑋𝑚(𝑡),Xε(0) = x0
Assume that the insurer maximizes the terminal T moment expected utility in the financial market, the portfolio
index utility, which takes the form of:
V(X(T)) = −
1
𝛾
𝑒−𝛾𝑋(𝑇)
Where 𝛾 > 0 is the ambiguity aversion coefficient of the insurer. The insurer's goal is to find the optimal
reinsurance-investment strategy𝜀∗
(𝑡) = (𝑞1
∗
(𝑡), 𝑞2
∗
(𝑡), 𝜋∗
(𝑡), β∗
(𝑡))to maximizing the expectation of end-use
wealth utility for insurers. The objective function of the insurer is:
Vε
(t, x, l, h) = E(u(Xε
(T))|Xε
(t) = x, L(t) = l)
The value function of the optimization problem is:
V(t, x, l, h) = Vε
(t, x, l, h)
ε∈Π
sup
, V(T,x,l,h)=v(x),
Assume that the ambiguity information is described by the probability P and the reference model is measured by
the probability 𝒫Φ
which is equivalent to P：𝒫: = *𝒫Φ
|𝒫Φ
~𝑃+.Next, the optional measure set is constructed,
defining the procedure :Φ(𝑡) = (Φ0(𝑡), Φ1(𝑡), Φ2(𝑡), Φ3(𝑡))s.t.:
1. Φ(𝑡) is a ℊ(𝑡)- measurable for any t[0,T]
2. Φi(t) = Φi(t, ω), i = 0,1,2,3 andΦi(t) ≥ 0 for all (t, ω) ∈ ,0, T- × Ω.
3.∫ ||Φ(t)||2
𝑑𝑡 < ∞;
𝑇
0
Let Σ be all processes shaped as Φ，for all Φ ∈ Σ define a G-adaptation process under a real measure
*ΛΦ
(𝑡)|𝑡 ∈ ,0, T-+.From Ito differentiation we have:
𝑑ΛΦ
(𝑡) = ΛΦ
(𝑡−)(−Φ0(𝑡)dW0 − Φ1(𝑡)𝑑𝑊1 − Φ2(𝑡)𝑑𝑊2 − (1 − Φ3(𝑡))𝑑Mp
)
Where ΛΦ
(0) = 1，P-a.s. ΛΦ
(𝑡) is a (P,G)- martingale，E[ΛΦ
(𝑇)] =1, for each Φ ∈ Σ,A new optional measure
is absolutely continuous to P, defined as
𝑑𝑃Φ
𝑑𝑃
|𝒢𝑇
= ΛΦ
(𝑇).
So far, we have constructed a class of real-world probability measures 𝑃Φ
，where Φ ∈ Σ，From Gisanova's
theorem[21]
,it follows that:
𝑑Wi
𝑃Φ
(𝑡) = dWi(𝑡) + Φ𝑖(𝑡)𝑑𝑡, 𝑖 = 0,1,2
Therefore the wealth process in the 𝑃Φ
is：
d𝑋𝜀
(𝑡) = ,xr + 𝜋(𝛼𝑙 − Φ1√𝑙) + β(1 − H(t))δ(1 − Δ)
+λμX(θ1 − η1) + λpμY((θ2 − η2) + η1𝑞1(𝑡)) + η2𝑞2(𝑡))
+√(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2Φ0-dt + 𝜋√𝑙𝑑W1
𝑃Φ
+β(1 − H(t−))ζdMp
+ √(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2𝑑W0
𝑃Φ
(2)
We assume that the insurer determines a robust portfolio strategy such that the worst-case scenario is the best
option. The insurer penalizes any deviation from the sub-reference model with a penalty that increases with this
deviation, using relative entropy to measure the deviation between the reference measure and the optional
measure. Inspired by Maenhout[19]
Branger[20]
the problem is modified to define the value function as：
V(t, x, l. h) = 𝐸𝑡,𝑥,𝑙,ℎ
𝑃Φ
0−
1
𝛾
𝑒−𝛾𝑋(𝑇)
+ ∫ 𝐺(𝑢, 𝑋𝜀(𝑢), 𝜙(𝑢))𝑑𝑢
𝑇
𝑡
1
Φ∈Σ
𝑖𝑛𝑓
𝜀∈𝛱
𝑠𝑢𝑝
𝐸𝑡,𝑥,𝑙,ℎ
𝑃Φ
Calculated under the optional measure , the initial value of the stochastic process is 𝑋𝜀(𝑡)=x, S(t)=s,
H(t)=h,
G(u, Xε(u), ϕ(u)) =
Φ0
2
2Ψ0(t, Xε(t), ϕ(t))
+
Φ1
2
2Ψ1(t, Xε(t), ϕ(t))
+
Φ2
2
2Ψ2(t, Xε(t), ϕ(t))
+
(Φ3 ln Φ3 − Φ3 + 1)hp
(1 − h)
2Ψ3(t, Xε(t), ϕ(t))
Where Ψ0 ≥ 0，Ψ1 ≥ 0，Ψ2 ≥ 0，Ψ3 ≥ 0 is state-dependent，let Ψ0 = −
𝑣0
𝛾V(t,x,l,h)
, Ψ1 = −
𝑣1
𝛾V(t,x,l,h)
, Ψ2 =
−
𝑣2
𝛾V(t,x,l,h)
, Ψ3 = −
𝑣3
𝛾V(t,x,l,h)
𝑣𝑖,i=0,1,2,3 is the risk aversion factor, the lager Ψi. According to the dynamic planning principle, the HJB
equation is as follows:
𝒜𝜀,𝜙
𝑉 + 𝐺(𝑢, 𝑋𝜀(𝑢), 𝜙(𝑢)) = 0
Φ𝜖Σ
𝑖𝑛𝑓
𝜀𝜖∏
𝑠𝑢𝑝
#(3)
𝒜𝜀,𝜙
is the infinitesimal operator under the measure 𝑃Φ
.

III. Robust optimal reinsurance-investment strategy solving
This section will solve the robust optimal problem constructed in the previous section. This paper
divides the value function into pre-default and post-default components according to the time of default of the
credit bond:
V(T, x, l, h) = {
V(T, x, l, 0)，h = 0(before default)
V(T, x, l, 1)，h = 1(after default)
By decomposing the value function into two sub-functions, denoted as the value function before the zero-
coupon bond default and the value function after the zero-coupon bond default, the two sub-HJB equations are
obtained and solved successively to obtain the reinsurance and risky asset investment strategies and value
function expressions after default, and the reinsurance, risky asset and credit bond investment strategies and
value function expressions before default.
1.5 Optimal reinsurance and investment decisions after default
When H (t)=1，τ ∧ T ≤ t ≤ T,the insurer has constituted a default at or before time t, the HJB equation
degenerates to：
𝑉𝑡 +
[𝑟𝑥 + 𝜋(𝛼𝑙 − Φ1√𝑙) + 𝜆μX(θ1 − η1) + λpμY(θ2 − η2) + 𝜆μXη1𝑞1(𝑡) + λpμYη2𝑞2(𝑡) +
√(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2Φ0]𝑉
𝑥 +
1
2
(π2
l + λ(q1ςX)2
+ λp(q2ςY)2
+ 2λpμXμY)Vxx + πlςρVxl +
(k(ω − l) − Φ1𝜍𝜌√𝑙 − Φ2𝜍√1 − 𝜌2√𝑙)Vl +
1
2
ς2
lVll −
Φ0
2
2v0
𝛾 −
Φ1
2
2v1
𝛾 −
Φ2
2
2v2
𝛾 = 0 (4)
Satisfied ：V(T, x, l, 1) = −
1
γ
e−γX
The solution can be assumed to be of the form:
V(t, x, y, l, 1) = −
1
γ
exp*−γ𝑒𝑟(𝑇−𝑡)
x + G(t, l)+, G(T, l) = 0#(5)
Taking each partial derivative of V:
{
𝑉𝑡 = [γrx𝑒𝑟(𝑇−𝑡)
+ Gt]V, 𝑉
𝑥 = −γ𝑒𝑟(𝑇−𝑡)
V
𝑉
𝑥𝑥 = γ2
𝑒2𝑟(𝑇−𝑡)
𝑉，𝑉𝑥𝑙 = −γ𝐺𝑙𝑉
𝑉𝑙 = 𝐺𝑙𝑉, 𝑉𝑙𝑙 = 𝐺𝑙𝑙𝑉 + 𝐺𝑙
2
𝑉
The minimum value point of Φ∗
is obtained from the first order condition as:
{
Φ0
∗
= v0er(T−t)
√(q1γ1)2 + (q2γ2)2 + 2ρ
̂q1q2γ1γ2
Φ1
∗
= v1er(T−t)
π√l, Φ2
∗
= −ςρ√lGlv2
1
γ
Bringing in the HJB equation yields:
γrx𝑒𝑟(𝑇−𝑡)
+ Gt − ,λμX(θ1 − η1) + λpμY(θ2 − η2)-
γ𝑒𝑟(𝑇−𝑡)
+ k(ω − l)𝐺𝑙 +
(𝜍𝜌√𝑙𝐺𝑙)
2
v2
2
+
1
2
𝜍2
𝑙(𝐺𝑙𝑙 + 𝐺𝑙
2
)
+ *𝑓2(𝜋, 𝑡)+
π
inf
+ {λγ𝑒𝑟(𝑇−𝑡)
𝑓1(𝑞1, 𝑞2, 𝑡)}
𝑞1,𝑞2
inf
= 0#(6)
Where
𝑓1(𝑞1, 𝑞2, 𝑡) = −,μXη1𝑞1(𝑡) + pμYη2𝑞2(𝑡)-
+
(γ + v0)𝑒𝑟(𝑇−𝑡)
2
,(𝑞1𝜍𝑋)2
+ p(𝑞2𝜍𝑌)2
+ 2pςXςY𝑞1𝑞2-
𝑓2(𝜋, 𝑡) = 𝜋𝑙𝑒𝑟(𝑇−𝑡)(ςρ𝐺𝑙(γ + v1) − 𝛼) +
1
2
(γ + v1)lπ2
Theorem III.1 Let m=
μX(η1𝜎𝑌
2−pη2μY
2)
𝜎𝑥
2𝜎𝑌
2−𝑝μX
2μY
2 , n=
μY(η2𝜎𝑋
2−η1μX
2)
𝜎𝑥
2𝜎𝑌
2−𝑝μX
2μY
2 ,then there exist 𝑡1, 𝑡2𝑡1
̂, 𝑡2
̂ and the following values
can be obtained:

𝑡1 = 𝑇 −
1
𝑟
ln
𝑚
γ + v0
, 𝑡2 = 𝑇 −
1
𝑟
ln
𝑛
γ + v0
𝑡1
̂ = 𝑇 −
1
𝑟
ln
μXη1
(γ + v0)p(𝜍𝑋
2 + μXμY)
𝑡2
̂ = 𝑇 −
1
𝑟
ln
μyη2
(γ + v0)(𝜍𝑌
2 + μXμY)
When m ≤ γ(n ≤ γ),let 𝑡2 = 𝑇(𝑡1 = 𝑇).When m > γ(n > γ),let 𝑡2 = 0(𝑡1 = 0)(1)If 𝑚 ≤ n,then 𝑞1
∗
≤ 𝑞2
∗
，for
all tϵ,0, T-，The reinsurance strategy corresponding to the problem is(𝑞1
∗
, 𝑞2
∗
) = {
(𝑞1
̂ , 𝑞2
̂)，0 ≤ 𝑡 ≤ 𝑡2
(𝑞1
̃, 1
̃)，𝑡2 ≤ 𝑡 ≤ 𝑡1
̂
(1,1)，𝑡1 ≤ 𝑡 ≤ 𝑇
(2)If
n > m,then for alltϵ,0, T-，The reinsurance strategy corresponding to the problem
is(𝑞1
∗
, 𝑞2
∗
) = {
(𝑞1
̂ , 𝑞2
̂),0 ≤ 𝑡 ≤ 𝑡1
(1, 𝑞2
̃), 𝑡1 ≤ 𝑡 ≤ 𝑡2
̂
(1,1), 𝑡2
̂ ≤ 𝑡 ≤ 𝑇
. Where
𝑞1
̂ =
μX(η1𝜍𝑌
2
− pη2μY
2)
(γ + v0)𝑒𝑟(𝑇−𝑡)(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
#(7)
𝑞2
̂ =
μY(η2𝜍𝑋
2
− η1μX
2)
(γ + v0)𝑒𝑟(𝑇−𝑡)(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
#(8)
𝑞1
̃ =
μXη1 − pμXμY(γ + v0)𝑒𝑟(𝑇−𝑡)
𝜍𝑋
2(γ + v0)𝑒𝑟(𝑇−𝑡)p
#(9)
𝑞2
̃ =
μyη2 − μXμY(γ + v0)𝑒𝑟(𝑇−𝑡)
𝜍𝑌
2(γ + v0)𝑒𝑟(𝑇−𝑡)
#(10)
Proof：By finding the first-order partial derivatives, second-order partial derivatives and second-order mixed
partial derivatives for𝑓1, the following system of equations and the Hessian array are obtained:
{
𝜕𝑓1
𝜕𝑞1
= −μXη1 + (γ + v0)𝑒𝑟(𝑇−𝑡)
,𝜍𝑋
2
𝑞1 + pςXςY𝑞2- = 0
𝜕𝑓1
𝜕𝑞2
= −μYη2 + (γ + v0)𝑒𝑟(𝑇−𝑡)
,𝜍𝑌
2
𝑞2 + ςXςY𝑞1- = 0
(11)
|
|
𝜕2
𝑓1
𝜕𝑞1𝜕𝑞1
𝜕2
𝑓1
𝜕𝑞1𝜕𝑞2
𝜕2
𝑓1
𝜕𝑞2𝜕𝑞1
𝜕2
𝑓1
𝜕𝑞2𝜕𝑞2
|
| = 𝑒4𝑟(𝑇−𝑡)
(γ + v0)2
|
−𝜍𝑥
2
(γ + v0) −pςXςY
−pςXςY −𝑝𝜍𝑦
2
(γ + v0)
|
From the Hessian array positive definite it is known that 𝑓1(𝑞1, 𝑞2, 𝑡) is a convex function and there exist
extreme value points; solving the system of equations (11) yields (7). (8).
Obviously，𝑞1
̂ 、𝑞2
̂ are both increasing functions with respect to t, when 𝑚 ≤ n,0 ≤ 𝑡 ≤ 𝑡1 or n > m，0 ≤
𝑡 ≤ 𝑡2the solution as (7)(8)，Also the values of 𝑡1 and 𝑡2 can be found. When 𝑚 ≤ n，𝑡2 ≤ 𝑡 ≤ 𝑡1
̂ ，then
(q1
∗
, q2
∗
) = (𝑞1
̃, 1),q1 ∈ ,0,1-.When 𝑛 > m，𝑡1 ≤ 𝑡 ≤ 𝑡2
̂ then (q1
∗
, q2
∗
) = (1, 𝑞1
̃), q2 ∈ ,0,1-,Separate solve
df1(t,q1,1)
dq1
= 0，
df1(t,1,q2)
dq2
= 0 , then can get 𝑞1
̃, 𝑞2
̃ as (9), (10) End.
Theorem III.2 The post-default insurer's optimal risky asset investment strategy is to
π∗
=
α − ςρ(γ + v1)𝐺1
#(12)
The expression of the optimal value function is:
V(t, x, y, l, 1) = −
1
γ
exp*−γ𝑒𝑟(𝑇−𝑡)(t)x + G1(t)l + G2(t)+#(13)
𝑤𝑕𝑒𝑟𝑒 ρ ≠ ±1, 𝐺1 =
𝑙1𝑙2 − 𝑙1𝑙2𝑒−
1
2
ς2(𝑙1−𝑙2)(γ+v1)(1−ρ2)(𝑇−𝑡)
𝑙2 − 𝑙1𝑒−
1
2
ς2(𝑙1−𝑙2)(γ+v1)(1−ρ2)(𝑇−𝑡)
, #(14)
ρ = 1, 𝐺1 =
α2
2(γ + v1)(ας + k)
(1 − 𝑒−(ας+k)(𝑇−𝑡)
), #(15)

ρ = −1, k ≠ ας, 𝐺1 =
α2
2(γ + v1)(𝑘 − ας)
(1 − 𝑒(ας−k)(𝑇−𝑡)
), #(16)
ρ = −1, 𝑘 = ας, 𝐺1 =
α2
2(γ + v1)
, #(17)
𝑙1,2 =
−(αςρ + k) ± √(αςρ + k)2 + α2ς2(1 − ρ2)
γς2(1 − ρ2)
G2 =
γ
r
,λ(μX(θ1 − η1) + λp(μY(θ2 − η2)-,1 − 𝑒𝑟(𝑇−𝑡)
-
+𝑘𝜔 ∫ G1(s)ds
𝑇
𝑡
+ ∫ γ𝑒𝑟(𝑇−𝑡)
𝑓1(𝑞1
∗
, 𝑞2
∗
, 𝑠)𝑑𝑠
𝑇
𝑡
#(18)
Proof: Let
G(t, l) = G1(t)l + G2(t)#(19)
Bringing equation (19) into equation (6) can be obtained as the following two equations:
G1
’
−
α2
2(γ + v1)
− G1(αςρ + k) +
1
2
(γ + v1)ς2
G1
2
(ρ2
− 1) = 0#(20)
G2‘ − ,λ(μX(θ1 − η1) + λp(μY(θ2 − η2)-γ𝑒𝑟(𝑇−𝑡)
+𝑘𝜔G1 + λγ𝑒𝑟(𝑇−𝑡)
𝑓1(𝑞1
∗
, 𝑞2
∗
, 𝑡) = 0#(21)
Solving equation (20) yields equations (14)-(17). From equation (21) we get (18), and the specific expression of
(18) is discussed in the following cases：
1)When 𝑚 ≤ n,0 ≤ 𝑡 ≤ 𝑡2(n > m, 0 ≤ 𝑡 ≤ 𝑡1)
G2 =
γ
𝑟
,λ(μX(θ1 − η1) + λp(μY(θ2 − η2)-,𝑒𝑟(𝑇−𝑡)
− 1- + G1(t)
̂
−λ(T − t)*μXη1
μX(η1𝜍𝑌
2
− pη2μY
2)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
+ pμYη2
μY(η2𝜍𝑋
2
− pη1μX
2)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
−
1
2
,(
μX𝜍𝑋(η1𝜍𝑌
2
− pη2μY
2)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
)
2
+ p (
μY𝜍𝑌(η2𝜍𝑋
2
− pη1μX
2)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
)
2
+2pςXςY
μX(η1𝜍𝑌
2
− pη2μY
2
)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
μY(η2𝜍𝑋
2
− pη1μX
2
)
(𝜍𝑥
2𝜍𝑌
2 − 𝑝μX
2μY
2)
-+
2）When 𝑚 ≤ 𝑛, 𝑡2 ≤ 𝑡 ≤ 𝑡1
̂,
G2 =
γ
r
− 1-
+
𝜆𝛾
(𝛾 + v0)
(
μXη1
𝜍𝑋𝑝
)
2
(
1
2
− p)(T − t) + λ(−
μXη1pμXμY
𝜍𝑋
2p
−λpμYη2 +
μXη1μXμY
𝜍𝑋
+ ςY
μXη1
ςX
)
γ
r
(𝑒𝑟(𝑇−𝑡)
− 1)
+,(
μXμY
𝜍𝑋
)
2
+ p𝜍𝑌
2
− 2𝑝ςY
μXμY
𝜍𝑋
-
λγ(𝛾 + v0)
2r
(𝑒2𝑟(𝑇−𝑡)
− 1) + 𝑘𝜔G1(t)
̂
3）When 𝑚 > 𝑛, 𝑡1 ≤ 𝑡 ≤ 𝑡2
̂,optimal reinsurance strategy (𝑞1
∗
, 𝑞2
∗
) =(1,𝑞2
̃),
G2 =
γ
r
− 1-
+
λγ
(γ + 𝜈0)
(
μyη2
𝜍𝑌
)
2
(
1
2p
− 1) (T − t) + 𝑘𝜔G1(t)
̂
+λμyη2 (μXη1 − 𝑝μYη2
μXμY
𝜍𝑌
2
− ςX
μyη2
𝜍𝑌
+
μyη2μXμY
2𝜍𝑌
2
)
γ(𝑒𝑟(𝑇−𝑡)
− 1)
r
+(𝜍𝑋
2
− 2pςX
μXμY
𝜍𝑌
+ p(
μXμY
𝜍𝑌
)2
)
λ𝛾(𝛾 + v0)
4r
(𝑒2𝑟(𝑇−𝑡)
− 1)
4) When 𝑚 > 𝑛, 𝑡2
̂ ≤ 𝑡 ≤ T(𝑚 ≤ 𝑛, 𝑡1
̂ ≤ 𝑡 ≤ 𝑇), optimal reinsurance strategy (𝑞1
∗
, 𝑞2
∗
) =(1,1),
G2 =
γ
r
,λμXθ1 + λpμYθ2-,𝑒𝑟(𝑇−𝑡)
− 1-
+𝑘𝜔G1(t)
̂ +
λγ(γ + v0)
4
,(𝜍𝑋)2
+ p(𝜍𝑌)2
+ 2pςXςY-(𝑒2𝑟(𝑇−𝑡)
− 1)
Where G1(t)
̂ = kω ∫ G1(s)ds
T
t
=
𝑙2𝑘𝜔(𝑇 − 𝑡) − 2𝑘𝜔(𝜍2
(1 − 𝜌2
)−1

∗ 𝑙𝑛|
𝑙1 − 𝑙2
𝑙1 − 𝑙2𝑒0.5𝜎2(𝑙1−𝑙2)(1−𝜌2)(γ+v1)(𝑇−𝑡)
|)，ρ ≠ ±1
α2
𝑘𝜔
2(𝑘 + ας)(γ + v1)
(T − t) +
1
2
(
α
(k + ας)(γ + v1)
)2
𝑘𝜔(1 − 𝑒(𝑘+ας)(𝑇−𝑡)
)，ρ = 1
α2
𝑘𝜔
2(𝑘 − ας)(γ + v1)
(T − t) +
1
2
(
α
(k − ας)(γ + v1)
)2
𝑘𝜔(1 − 𝑒(𝑘−ας)(𝑇−𝑡)
)，ρ = −1, k ≠ ας
𝑘𝜔
,(T−t)α-2
4(γ+v1)
，ρ = −1，k = ας. End.
1.6 Optimal reinsurance and investment decisions before default
This section considers the optimal pre-default reinsurance-investment strategy and the value function expression
based on the previous section, when H(t)=0, 0 ≤ t ≤ τ ∧ T.Let the solution of the default prior value function
have the following form:
V(T, x, l, 0) = −
1
𝛾
𝑒𝑥𝑝{−𝛾𝑒𝑟(𝑇−𝑡)(𝑡)𝑥 + 𝐾(𝑡, 𝑙)}#(22)
Satisfying the boundary conditions V(T, x, l, 0)=V(x), K(T, l)=0, the HJB equation is transformed into:
𝑉𝑡 + ,𝜋(𝛼𝑙 − Φ1√𝑙) + βδ(1 − Δ) + λμX(θ1 + η1𝑞1(𝑡) − 1))
+λpμY .θ2 + η2(𝑞2(𝑡) − 1)√(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2Φ01 𝑉
𝑥
+
1
2
((𝜋)2
𝑙 + λ(𝑞1𝜍𝑋)2
+ λp(𝑞2𝜍𝑌)2
+ 2λp𝜍𝑋𝜍𝑌)𝑉
𝑥𝑥
+𝜋𝑙𝜍ρ𝑉𝑥𝑙 + (k(ω − l) − Φ1𝜍𝜌√𝑙 − Φ2𝜍√1 − 𝜌2√𝑙))Vl +
1
2
ς2
lVll
+𝑉(𝑒𝛾β(t)ζ𝐻1(𝑡)+𝐾(𝑡,𝑙)−G(𝑡,𝑙)
− 1)𝑕𝑝
−
Φ0
2
2v0
𝛾 −
Φ1
2
2v1
𝛾 −
Φ2
2
2v2
𝛾 −
(Φ3 ln Φ3 − Φ3 + 1)hp
(1 − h)
2v2
𝛾 = 0
Similarly find the partial derivative of V:
{
𝑉𝑡 = [γ𝑒𝑟(𝑇−𝑡)
+ Kt’]V, 𝑉
𝑥 = −γ𝑒𝑟(𝑇−𝑡)
V
𝑉
𝑥𝑥 = γ2
𝑉，𝑉𝑥𝑙 = −γ𝑒𝑟(𝑇−𝑡)
𝐾𝑙𝑉
𝑉𝑙 = 𝐾𝑙𝑉, 𝑉𝑙𝑙 = 𝐾𝑙𝑙𝑉 + 𝐾𝑙
2
𝑉
Bringing the above expression into the HJB equation and fixing the reinsurance-investment strategy, the
minimum point of Φ is obtained according to the first-order condition as:
{
Φ0
∗
= v0𝑒𝑟(𝑇−𝑡)
√(𝑞1𝛾1)2 + (𝑞2𝛾2)2 + 2ρ
̂𝑞1𝑞2𝛾1𝛾2
Φ1
∗
= v1𝑒𝑟(𝑇−𝑡)
𝜋√𝑙, Φ2
∗
= −𝜍𝜌√𝑙K𝑙v2
1
𝛾
Φ3
∗
= 𝑒𝑥𝑝*
v3
𝛾
.𝑒−𝛾β(t)ζ𝑒𝑟(𝑇−𝑡)+𝐺1(𝑡,𝑙)−G(𝑡,𝑙)
− 1/+
#(23)
After bringing (23) into the HJB equation, by inverting the investment strategy we can obtain:
π∗
=
α − ςρ(γ + v1)Kl
, β∗
=
ln
1
∆Φ3
+ K(t, l) − G(t, l)
ζγ𝑒𝑟(𝑇−𝑡)
and obviously the expressions for the reinsurance strategy before and after the bond default are the same, so that
𝑔1(𝑡, q1, q2) = 𝑓1(𝑡, q1, q2) = −,μXη1q1 + pμYη2q2-
+
1
2
,(q1ςX)2
+ p(q2ςY)2
+ pςXςYμXμYγ2
-(γ + v0)𝑒𝑟(𝑇−𝑡)
Bringing ε∗
= (π∗
, β∗
, 𝑞1
∗
, 𝑞2
∗
) into the equation, after finishing, we get:
,−γ𝑟𝑒𝑟(𝑇−𝑡)
+ Kt’- − ,λμX(θ1 − η1)
+λpμY(θ2 − η2)-γ𝑒𝑟(𝑇−𝑡)
+ k(ω − l)Kl +
1
2
ς2
l(Kll + Kl
2
)
−λγer(T−t)
𝑔1(q1
∗
, q2
∗
, t) − g2(π∗
, t) − g3(β∗
, t) = 0#(24)
Where

𝑔2(π∗
, t) = π∗
𝑙𝑒𝑟(𝑇−𝑡)(ςρ𝐾𝑙(γ + v1) − 𝛼) +
1
2
(γ + v1)lπ∗2
𝑒2𝑟(𝑇−𝑡) (25)
𝑔3(β∗
, t) = −
(Φ3 − 1)γhp
v3
− δβγer(T−t)
#(26)
The derivative of equation (26) with respect to β ：
β∗
=
ln
1
∆Φ3
+ K(t, l) − G(t, l)
ζγ𝑒𝑟(𝑇−𝑡)
#(27)
Bringing it into Φ3
∗
get :
hp
v3
Φ3(𝑡) ln Φ3(𝑡) + hp
Φ3(𝑡) −
𝛿
𝜁
= 0. The equation has dimension one positive roots
Φ3.
Let K(t, l)=K1(t)l + K2(t)，satisfy the boundary conditions K1(T)=0,K2(T) = 0，bring it to HJB equation(24)，
eliminating the effect of l on the equation yields two equations:
K1
’
− k𝐾1 +
1
2
ς2
𝐾1
2
−
𝐾1 − 𝐺1
ζ
δ − (ςρ𝐾1 + α(δ − r))α(δ − r)
+
1
2
(ςρ𝐾1 + α(δ − r))
2
− (ςρ𝐾1 + α(δ − r))ςρ𝐾1 = 0#(28)
K2
’
− ,λμX(θ1 − η1) + λpμY(θ2 − η2)-γ𝑒𝑟(𝑇−𝑡)
+ λpςXςYγ2
(𝑒𝑟(𝑇−𝑡)
)2
+kω𝐾1 −
ln
1
∆Φ3
+ 𝐾2 − 𝐺2
ζ
δ + (1 −
1
∆Φ3
) hp
− f(t, 𝑞1
∗
, 𝑞2
∗) = 0#(29)
When ρ ≠ ±1,equation (28) is the first-order RICCATI equation that，from the existence uniqueness of the
solution 𝐾1 = 𝐺1.Then we solve equation(29):let I(t)=𝐾2 − 𝐺2, Satisfying the boundary condition I(T) = 0, then
using (29) and (21) we get：I’ = K2
’
− G2
’
=
δ
ζ
I +
δ
ζ
ln
1
∆Φ3
+
γ(Φ3−1)
v3
hp
,
I = (ln
1
∆Φ3
+ ∆ − 1)𝑒
−
δ
ζ
(T−t)
− ln
1
∆Φ3
− ∆ + 1.Thus ：
𝐾2 = (ln
1
∆Φ3
+
γ∆(Φ3 − 1)
v3
) 𝑒
−
δ
ζ
(T−t)
− ln
1
∆Φ3
+
γ∆(Φ3 − 1)
v3
+ 𝐺2(30)
Theorem III.3 The optimal pre-default investment and bond investment strategies are:
π∗
=
α − ςρ(γ + v1)Kl
, 𝛽∗
=
𝑣3ln
1
ΔΦ3
+ Δ𝛾(Φ3 − 1)(𝑒
−
𝛿
𝜁
(T−t)
− 1)
𝑣3𝜁𝛾𝑒𝑟(𝑇−𝑡)
The expression of the value function before default is:
𝑉(𝑡, 𝑥, 𝑙, 0) = −
1
𝛾
exp{−𝛾𝑒𝑟(𝑇−𝑡)
𝑥 + K1(𝑡)𝑙+K2(𝑡)+
Where K1(t) = G1(t)，as the formula (14)-(17)，K2(t) as the formula (30).
IV. Parameter Sensitivity analysis
In this chapter, we give several numerical examples to test the effect of model parameters on the
optimal strategy. First, to make the parameters more realistic, the model parameters for credit bonds are set as
follows, based on the estimates from Berndt (2008)[22]
and Collin-Dufrensn & Solnik (2001)[23],
which are
estimated from market data:1/Δ = 2.53, hQ
= 0.013, 𝜁 = 0.52.The claim amounts for the first and second
categories of insurance respectively meet the parameters of λ𝑋 = 1.5, λ𝑌 = 1exponential distribution . If no
special statement is made, other model parameters are assumed as follows:： t = 3, T = 10, 𝑟 = 0.06, 𝛾 =
4，η1 = 2, η2 = 3, p = 0.5, 𝜈0 = 𝜈1 = 𝜈3 = 1, ρ = 1, ς = 0.16, 𝑘 = 2, α = 1.5.
1.7 Influence of parameters on optimal reinsurance strategy
Figure.1-2 represents the variation of the optimal reinsurance strategy with the parameters. where Figure.1
shows the variation of reinsurance strategy with probability p. When p increases, the insurer will reduce the
amount of reinsurance retention for class I reinsurance and increase the amount of reinsurance retention for class
II reinsurance. figure.2 represents the effect of risk aversion coefficient 𝛾 on reinsurance strategy, when the risk
aversion coefficient is larger, the insurer will reduce the reinsurance strategy and purchase more reinsurance to
diversify Claims risk.

Figure. 1 Figure. 2
1.8 Influence of parameters on optimal investment strategy
Where Figure.3 represents the effect of the risk-free rate r on the optimal investment strategyπ∗
, when the risk-
free rate increases, the insurer will invest less in risky assets and more assets in risk-free assets. Figure.4
represents the effect of the volatile reversion rate k on the optimal investment strategyπ∗
. When the reversion
rate increases, the insurer will increase its investment in risky assets.
Figure. 3 Figure. 4
Figure.5 shows the impact of risk premium
1
Δ
on credit bonds, when the risk premium increases, investment in
credit bonds is increased. Figure.6 indicates that when the default loss rate 𝜁 increases, insurers will invest less
in credit bond.
Figure. 5 Figure. 6
1.9 Influence of ambiguity aversion coefficient on strategy
Figure.7 to Figure.9 represent the effect of optimal reinsurance-investment strategy subject to ambiguity
aversion sparsity. It can be seen that as the ambiguity aversion coefficient increases, the insurer's uncertainty
about the model increases and therefore reduces its optimal reinsurance-investment strategy.
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6
p
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
q1
q2
2 2.5 3 3.5 4 4.5 5 5.5 6
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
q1
q2
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
r
0.05
0.1
0.15
0.2
0.25
0.3
1 1.5 2 2.5 3 3.5 4 4.5 5
k
0.222
0.2225
0.223
0.2235
0.224
0.2245
0.225
0.2255
0.226
1 2 3 4 5 6 7 8
1/
0
0.5
1
1.5
2
2.5
3
3.5
4
1
, =1.5
2
, =2.53
3
, =5
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.5
1
1.5
2
2.5
3
3.5
1
,1/ =1.5
2
,1/ =2.53
3
,1/ =5

Figure. 7 Figure. 8
Figure. 9
V. Conclusion
This paper assumes that the insurer owns two types of insurance business with sparse dependence risk,
and the claim process is described by a diffusion approximation model, and secondly, the insurer expands its
investment types by investing in the financial market with a stock, a risk-free asset and a credit bond, with the
stock price described by a Heston model and the credit bond price described by an approximate model. The
credit bond price is described by the approximate model, and considering that the insurer is ambiguous averse,
the robust optimal reinsurance-investment problem is established, and the explicit expressions of the robust
optimal reinsurance-investment and optimal value function are obtained by using stochastic control theory,
dynamic programming principle, and HJB equation, and sensitivity analysis is performed on the model
parameters. Based on this paper, further discussions can be made: 1) other situations of the claim process can be
considered: such as jump diffusion or common shock. 2) time lag effects can be considered. 3) game problems
can be considered.
References
[1]. Browne S . Optimal investment policies for a firm with a random risk process: Exponential utility and
minimizing the probability of ruin[J]. Mathematics of Operations Research, 1995, 20(4):937-958.
[2]. Zhang Y , Zhao P , Teng X , et al. Optimal reinsurance and investment strategies for an insurer and a
reinsurer under Hestons SV model: HARA utility and Legendre transform[J]. Journal of Industrial and
Management Optimization, 2017, 13(5).
[3]. Zhu S , Shi J . Optimal Reinsurance and Investment Strategies Under Mean-Variance Criteria: Partial
and Full Information[J]. Journal of Systems Science and Complexity: English Edition, 2022, 35(4):22.
[4]. LI, Danping, RONG, et al. Optimal Investment Problem for an Insurer and a Reinsurer[J]. Journal of
Systems Science and Complexity: English Edition.,2015(6):18.
[5]. A D L , B X R A , A H Z . Time-consistent reinsurance–investment strategy for an insurer and a
reinsurer with mean–variance criterion under the CEV model[J]. Journal of Computational and Applied
Mathematics, 2015, 283:142-162.
[6]. Chun-Xiang A , Gu A L , Shao Y . Optimal Reinsurance and Investment Strategy with Delay in
Heston's SV Model[J]. Journal of the Chinese Society of Operations Research (English), 2021, 9(2):27.
[7]. Xiaoqin Gong, Shixia Ma, Qing Huang,, Robust optimal investment strategies for insurance and
reinsurance companies under stochastic interest rates and stochastic volatility (in English) [J]. Journal
of Nankai University: Natural Science Edition, 2019, 52(6):11.
[8]. Grandell J . Aspects of Risk Theory[M]. World Publishing Co. 1991.
[9]. Zhao H , Rong X , Zhao Y . Optimal excess-of-loss reinsurance and investment problem for an insurer
with jump-diffusion risk process under the Heston model[J]. INSURANCE -AMSTERDAM-, 2013.
[10]. Ceci C , Colaneri K , Cretarola A . Optimal reinsurance and investment under common shock
dependence between financial and actuarial markets[J]. Insurance: Mathematics and Economics, 2022,
105.
[11]. Zhang P . Optimal excess-of-loss reinsurance and investment problem with thinning dependent risks
under Heston model[J]. Journal of Computational and Applied Mathematics, 2021, 382(1).
[12]. Li S , Qiu Z . Optimal Time-Consistent Investment and Reinsurance Strategies with Default Risk and
Delay under Heston's SV Model[J]. Mathematical Problems in Engineering, 2021, 2021(1):1-36.
[13]. Zhu G . Time-consistent non-zero-sum stochastic differential reinsurance and investment game under
default and volatility risks[J]. Journal of Computational and Applied Mathematics, 2020, 374.
[14]. Zhenlong Chen, Weijie Yuan, Dengfeng Xia. Optimal Reinsurance-Investment Strategy with Default
Risk Based on Heston's SV Model [J]. Business Economics and Management, 2021, 000(005):56-70.
0 1 2 3 4 5 6 7 8 9 10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
q1
q2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
1
0.1
0.15
0.2
0.25
1 1.5 2 2.5 3 3.5 4
3
0.3442
0.34425
0.3443
0.34435
0.3444
0.34445
0.3445
0.34455
0.3446
1

[15]. Bo Y , Li Z , Viens F G , et al. Robust optimal control for an insurer with reinsurance and investment
under Heston's stochastic volatility model[J]. Insurance: Mathematics and Economics, 2013,
53(3):601-614.
[16]. Hui Meng, Li Wei, Ming Zhou. Robust reinsurance strategies for insurers under ambiguous aversion
[J]. Science of China:Mathematics, 2021, 51(11):28.
[17]. Zhang Y , Zhao P . Robust Optimal Excess-of-Loss Reinsurance and Investment Problem with Delay
and Dependent Risks[J]. Discrete Dynamics in Nature and Society, 2019, 2019(2):1-21.
[18]. Bielecki T R , Jang I . Portfolio optimization with a defaultable security[J]. Kluwer Academic
Publishers-Plenum Publishers, 2007(2).
[19]. Maenhout P J . Robust Portfolio Rules and Asset Pricing[J]. Review of Financial Studies, 2004(4):4.
[20]. Branger N , Larsen L S . Robust portfolio choice with uncertainty about jump and diffusion risk[J].
Journal of Banking & Finance, 2013, 37(12):1397-1411.
[21]. Shreve S E . Stochastic Calculus for Finance ll[M]. World Book Publishing, 2007.
[22]. Berndt A , Douglas R , Duffie D , et al. Measuring Default Risk Premia from Default Swap Rates and
EDFs[C]// Bank for International Settlements. Bank for International Settlements, 2005:1–18.
[23]. Collin-Dufresne P , Solnik B . On the Term Structure of Default Premia in the Swap and LIBOR
Markets[J]. Journal of Finance, 2001, 56(3):p.1095-1116.

�ݺ�ߣ

Robust Optimal Reinsurance and Investment Problem with p-Thinning Dependent and Default risks

More Related Content

Robust Optimal Reinsurance and Investment Problem with p-Thinning Dependent and Default risks