Systems and methods for risk simulation based transaction routing
Embodiments described herein further provide a routing server for routing incoming transaction requests based on a stochastic simulation of future risk level, e.g., demand flows such as required network resource, and/or the like.
1 . A routing system for routing a network access request based on a risk level, the routing system comprising:
a communication interface receiving the network access request comprising information relating to a network application including a plurality of application attributes;
a memory that stores a plurality of processor-executable instructions;
one or more processors that reads and executes the plurality of processor-executable instructions from the memory to perform operations comprising:
decomposing the network application into a set of discrete execution processes, each of which is subject to a respective stochastic entry time and a respective stochastic exit time in and out of the network application, and a respective stochastic waiting time having a hazard-based duration, respectively,
wherein the respective stochastic entry time or the respective stochastic exit time takes a form of a hazard rate model subject to an exponential process with time-varying coefficients;
a hardware-based stochastic simulator that:
conducting a stochastic simulation of entry dynamics or exit dynamics of the set of discrete execution processes based at least in part on simulating a stochastic entry intensity rate or an exit intensity rate according to a hazard rate model, based on the plurality of application attributes; and
generating, from simulation outputs, a predicted risk level associated with a forecasted demand flow from the network application over a future period;
wherein the one or more processors further perform operations comprising generating a routing decision for the network access request based on the predicted risk level.
2 . The routing system of claim 1 , wherein the operations further comprise:
causing a display of a visualization of the predicted risk level over a period of time on a user interface;
receiving, via the user interface, a user input that continuously move a marker on the user interface to indicate a changed time period; and
causing an updated display of a visualization of an updated predicted risk level over the changed time period on the user interface.
3 . The routing system of claim 1 , wherein the operations of generating the routing decision further comprising:
declining the network access request when the predicted risk level indicates a high risk; and
blocking future network access requests from the requestor server or relating to the target server.
4 . The routing system of claim 1 , wherein the operations of generating the routing decision further comprising:
routing, by a routing lookup table, the network access request to a destination server when the predicted risk level indicates an acceptable risk.
5 . The routing system of claim 1 , wherein the network access request comprises a transaction request from a requester server to initiate a transaction with a destination server, and wherein the network application comprises an investment vehicle including a plurality of vehicle attributes.
6 . The routing system of claim 5 , wherein the plurality of vehicle attributes include any combination of a strategy of the investment vehicle, a total vehicle life, an investment period length, a total commitment.
7 . The routing system of claim 5 , wherein the respective stochastic entry time or the respective stochastic exit time takes a form of a hazard rate model subject to an exponential process with time-varying coefficients.
8 . The routing system of claim 5 , further comprising:
calibrating, using a regression model, one or more entry intensity parameters or exit intensity parameters including a lagged intensity value, an age of the investment vehicle, and a risk-free rate and a market return based on historical capital call intensity data or historical distribution intensity data.
9 . The routing system of claim 5 , wherein the stochastic simulation includes simulating a capital call intensity rate or a distribution intensity rate for the investment vehicle, based at least in part on:
(i) obtaining a set of economic factors that drive capital calls, distributions or returns of a set of discrete portfolio investment opportunities decomposed from the investment vehicle;
(ii) generating the simulated capital call intensity rate or the simulated distribution intensity rate based on the set of economic factors;
(iii) generating simulated returns from the set of discrete portfolio investment opportunities by drawing the set of discrete portfolio investment opportunities from a Gaussian distribution;
(iv) generating a simulated entry time and a simulated exit time for each discrete portfolio investment opportunity based on transition probabilities computed from the simulated capital call intensity rate and the simulated distribution intensity rate;
(v) aggregating simulated activities of the set of discrete portfolio investment opportunities to obtain a capital drawdown, a distribution and a return value for the private equity investment vehicle; and
(vi) generating a distribution of state variable trajectory corresponding to investment states of the set of discrete portfolio investment opportunities.
10 . The routing system of claim 9 , wherein the simulated entry time and the simulated exit time are generated by sequentially simulating a set of investment states of each discrete portfolio investment opportunity,
wherein the set of investment states includes an uninvested state that transits to an invested state according to a first transition probability computed based on the simulated capital call intensity rate, and
wherein the invested state transits to a divested state according to a second transition probability computed based on the simulated distribution intensity rate.
11 . A method for routing a network access request based on a risk level, the method comprising:
receiving by a communication interface, the network access request comprising information relating to a network application including a plurality of application attributes;
decomposing, by a processor, the network application into a set of discrete execution processes, each of which is subject to a respective stochastic entry time and a respective stochastic exit time in and out of the network application, and a respective stochastic waiting time having a hazard-based duration, respectively,
wherein the respective stochastic entry time or the respective stochastic exit time takes a form of a hazard rate model subject to an exponential process with time-varying coefficients;
conducting, by a hardware-based stochastic simulator, a stochastic simulation of entry dynamics or exit dynamics of the set of discrete execution processes based at least in part on simulating a stochastic entry intensity rate or an exit intensity rate according to a hazard rate model, based on the plurality of application attributes;
generating, from simulation outputs, a predicted risk level associated with a forecasted demand flow from the network application over a future period; and
generating a routing decision for the network access request based on the predicted risk level.
12 . The method of claim 11 , further comprising:
causing a display of a visualization of the predicted risk level over a period of time on a user interface;
receiving, via the user interface, a user input that continuously move a marker on the user interface to indicate a changed time period; and
causing an updated display of a visualization of an updated predicted risk level over the changed time period on the user interface.
13 . The method of claim 11 , wherein the generating the routing decision further comprises:
declining the network access request when the predicted risk level indicates a high risk; and
blocking future network access requests from the requestor server or relating to the target server.
14 . The method of claim 11 , wherein the generating the routing decision further comprises:
routing, by a routing lookup table, the network access request to a destination server when the predicted risk level indicates an acceptable risk.
15 . The method of claim 1 , wherein the network access request comprises a transaction request from a requester server to initiate a transaction with a destination server, and wherein the network application comprises an investment vehicle including a plurality of vehicle attributes.
16 . The method of claim 15 , wherein the plurality of vehicle attributes include any combination of a strategy of the investment vehicle, a total vehicle life, an investment period length, a total commitment.
17 . The method of claim 15 , wherein the respective stochastic entry time or the respective stochastic exit time takes a form of a hazard rate model subject to an exponential process with time-varying coefficients.
18 . The method of claim 15 , further comprising:
calibrating, using a regression model, one or more entry intensity parameters or exit intensity parameters including a lagged intensity value, an age of the investment vehicle, and a risk-free rate and a market return based on historical capital call intensity data or historical distribution intensity data.
19 . The method of claim 15 , wherein the stochastic simulation includes simulating a capital call intensity rate or a distribution intensity rate for the investment vehicle, based at least in part on:
(i) obtaining a set of economic factors that drive capital calls, distributions or returns of a set of discrete portfolio investment opportunities decomposed from the investment vehicle;
(ii) generating the simulated capital call intensity rate or the simulated distribution intensity rate based on the set of economic factors;
(iii) generating simulated returns from the set of discrete portfolio investment opportunities by drawing the set of discrete portfolio investment opportunities from a Gaussian distribution;
(iv) generating a simulated entry time and a simulated exit time for each discrete portfolio investment opportunity based on transition probabilities computed from the simulated capital call intensity rate and the simulated distribution intensity rate;
(v) aggregating simulated activities of the set of discrete portfolio investment opportunities to obtain a capital drawdown, a distribution and a return value for the private equity investment vehicle; and
(vi) generating a distribution of state variable trajectory corresponding to investment states of the set of discrete portfolio investment opportunities.
20 . A non-transitory processor-readable storage medium storing a plurality of processor-executable instructions for analyzing or constructing an asset portfolio by simulating private equity cash flows, the processor-executable instructions being executed by a processor to perform operations comprising:
receiving, a communication interface, the network access request comprising information relating to a network application including a plurality of application attributes;
decomposing, by a processor, the network application into a set of discrete execution processes, each of which is subject to a respective stochastic entry time and a respective stochastic exit time in and out of the network application, and a respective stochastic waiting time having a hazard-based duration, respectively,
wherein the respective stochastic entry time or the respective stochastic exit time takes a form of a hazard rate model subject to an exponential process with time-varying coefficients;
conducting, by a hardware-based stochastic simulator, a stochastic simulation of entry dynamics or exit dynamics of the set of discrete execution processes based at least in part on simulating a stochastic entry intensity rate or an exit intensity rate according to a hazard rate model, based on the plurality of application attributes;
generating, from simulation outputs, a predicted risk level associated with a forecasted demand flow from the network application over a future period; and
generating a routing decision for the network access request based on the predicted risk level.