Field of the Invention

The present invention relates to digital system and platforms flexible adaptable by a plurality of users providing automated user-specific adaptable, flexible data processing, inter alia applicable to automated pricing of risk-transfers for an individual policy. In particular, it relates to digital system relying on Markov Chain modelling structures capturing states and/or state transitions and/or cashflows specific to a risk-transfer relying on mortality or life measures.

Background of the Invention

In traditional life risk-transfer industry, a finite-state Markov chain is often used as technical structure to capture and parametrize, i.e. to represent different state of the insured during data processing. Associating monetary amount vales with sojourn times in states and transitions between states allows to generate and forecast a present value for life risk-transfer losses to be expected within an associated error margin at a set interest rate value. Traditionally these model structures have been used with deterministic interest rate parameter settings and deterministic transition rate parameter settings, however in recent life risk-transfer forecast modelling and simulations, stochastic modelling structures of the interest parameters and transition rate parameters has gained attention in automation and data processing. This is of particular technical interest either if the risk (herein understood as the technical value of the forecasted frequency and aggregated event strength (actually occurring event loss) within a future measuring time window) associated with changes in the underlying interest parametrization and transition rate parameters is to be captured and forecasted, and/or if one wants to hedge this forecasted risk by building up balancing securities based on these same underlying rate parameter settings. A technical life risktransfer setup with stochastic transition rate settings is, where the underlying rate parameters are processed and captured by a finite-state Markov chain model structure and, in particular, dependence between the rate parameters is enabled within the structure. In the prior art systems, there are various technical data processing treatment of stochastic interest rates applied in life risk-transfer modelling and for forecasting stochastic mortality rate parameter values. Also combined model structures for stochastic interest and mortality rate parameters settings are known. Common to several of the prior art systems is that interest and mortality rate parameters are technically propagated and thus modelled by engineering model structures representing affine model processes for the data processing. This class of processes leads to technically tractable modelling and forecasts, where the core of the data processing is reduced to solve a system of ordinary differential equations instead of partial differential equations in order to find expected present values, which is of great technical advantage for data processing by finite state machines being based on finite state processors. The present invention technically allows for the application of affine processes in more general life risk-transfer modelling structures, thus making it easier and technically more efficient to forecast by data propagation and simulation expected present values.

It is known in the prior art systems, that finite-state Markov chain structures can be applied for data propagation in modelling risk-transfer and appropriate pricing parameters for a specific risk-transfer but can also be applied for forecasting of credit risks. Credit risk is defined herein as the actual and/or expected degree of value fluctuations in debt instruments and derivative structures due to measured changes in the underlying credit quality of associated counterparties or counter units. While vector autoregression (VaR) model structure applied to daily market risk calculations is able to generate about 250 forecast in one year, credit risk model structures technically are limited to about one forecast per year due to their tail structure. Thus, it would take a very long time to produce sufficient observation measurements for reasonable tests and weighting of forecast accuracy for these data processing structures. In addition, due to the nature of credit risk measurements, only a limited amount of historical data on credit losses can be assessed which is not enough to span several macroeconomic or credit cycle processing. These data limitations create a serious technical difficulty for the validation and calibration of credit risk model structures. It is to be mentioned that vector autoregression model structures are statistical based structures used to capture and propagate the relationship between multiple quantities and measures as they change over time. Thus, all VAR structures are types of stochastic process modelling. VAR modelling technically generalize the single-variable (univariate) autoregressive modelling by allowing for capturing multivariate time series. Similar to the autoregressive model structures, each measured or captured variable has a relation (function or equation) modelling and propagating its evolution over time. This relation includes the variable's lagged(past) values, the lagged values of the other variables in the model structure, and an error term. VAR model structures have the advantage that they do not require as much information about the forces influencing a variable as do structural model structures based on simultaneous equations. The only information, technically required, is an appropriate filter for variables which can be hypothesized to affect each other over time.

Thus, a major technical impediment to model validation (so called "back testing"), not only for life-risk-transfer modelling but also for credit risks modelling, is the small number of forecasts available with which to calibrate a model's forecast processing accuracy. There is a need to provide a more flexible modelling and simulation structure which can be used to forecast and propagate the relevant parameter values reliably to a future time range. Also basic credit risk model structures can typically be captured by a two-state Markov chain, where a jump from the initial state represents a default. An extension of this model, known in the prior art, is to let the default transition rate be modelled as a stochastic process itself such that it is possible for it to be dependent on the interest rate and other economic factors. This approach can be applied to various Markov chain model structures. In a more general treatment of the Markov chain approach to credit risk modelling with stochastic transition rates it can be shown how prices generally satisfy a system of partial differential equations. In both approaches, it can be shown how one can benefit from affine stochastic processes as transition intensities and economic measuring factors. If the modelling structure is particularly simple, the Riccati relation can be used to reduce the modelling problem of solving a system of partial differential equations to that of solving a system of ordinary differential equations, which can have significant technical advantages in terms of data processing efficiency and used time. The technical approach of the present invention has, inter alia, the advantage, that it allows to generalize these prior art methods allowing for risk-transfer pricing in more general decrement Markov chain model structures. Therefore, while Markov Chain modelling structures are used widely in risk-transfer technology and industries, its combination with configuration elements according to the present invention allows a level of flexibility and dynamism. The present inventive system can, inter alia, be applied for doubly stochastic Markov chain structures, in particular in life risk-transfer and credit risk forecast and simulations. It can further be applied for the generation of transition probabilities in certain doubly stochastic Markov chains, wherein the resulting output-signaling can e.g. be used for valuation of life insurance contracts.

Summary of the Invention

It is an object of the invention to provide more efficient and more accurate and more flexible systems and methods allowing newly dynamic and automated propagation of life risk parameters and associated automated dynamic risk-transfer pricing and dynamic premiums generation and assessment for a policy capturing the individual risk-transfer settings and parameter values. It is further an object of the present invention to allow for using a combination of Markov Chain modelling structures and configurable elements (States, State transitions, Cashflows) specific to a risktransfer, it instantiates, which was not possible by the prior art systems.

According to the present invention, these objects are achieved particularly through the features of the independent claims. In addition, further advantageous embodiments follow from the dependent claims and the description.

According to the present invention, the abovementioned objects are particularly achieved by the digital system, automated platform and automated method in that for life risk parameter propagation and for dynamic and automated risktransfer pricing by processing a plurality of individual risk-related parameters associated with a portfolio of risk-transfers of risk-exposed individuals, wherein each risk-transfer held associated with the portfolio is set by risk-transfer parameters of a risk-transfer policy defining the individual risk-transfer, wherein a combination of a Markov Chain modelling structure with configurable elements are applied at least comprising states and/or state transitions and/or cash-flows specific to the risk-transfer, and wherein a stochastic Markov data processing is applied to the Markov chain structure over a sequence of possible events in which the probability value of each event depends only on the state attained in the previous event, in that the digital system comprises a calculation engine comprising a data structure for capturing and/or storing deterministic transition and interest rate parameter values of a state discrete process to conduct data processing per policy, wherein parts of the process are used with stochastic transition and interest rates of a discrete process for conducting simulations on portfolio level, in that the digital system comprises an interface to the Markov chain structure, wherein for the stochastic Markov data processing, interest parameter values and transition rate parameter values are user-specific and flexible configurable and/or selectable from an associated digital library, and wherein within the stochastic Markov data processing setup and stochastic transition rates, these underlying rates are processed and modelled by the finite-state Markov chain structure, and in that for the data processing by the finite-state Markov chain structure, one or more transition functions are configurable via the data interface and/or selectable from the digital library, the transition functions linking at least two states within the Markov chain structure wherein all states of the Markov chain structure are linked to a antecedent and a successive state providing the data processing over the whole configurable Markov chain structure.

The technical approach of the present invention has, inter alia, the advantage, that it allows to generalize the above discussed prior art methods allowing for risk-transfer pricing in more general decrement Markov chain model structures. Therefore, while Markov Chain modelling structures are used widely in risk-transfer technology and industries, its combination with configuration elements according to the present invention allows a level of flexibility and dynamism. The present inventive system can, inter alia, be applied for doubly stochastic Markov chains, in particular in life risk-transfer and credit risk modelling. It can further be applied for the generation of transition probabilities in certain doubly stochastic Markov chains, wherein the resulting output-signaling can e.g. be used for valuation of life insurance contracts. Brief Description of the Drawings

The present invention will be explained in more detail below relying on examples and with reference to these drawings in which:

Figure 1 shows a block diagram illustrating schematically exemplary the relationships between the configuration and the calculation, as well as the hierarchy of the configurable elements themselves.

Figures 2 to 4 show block diagrams illustrating schematically an exemplary overview of slices with figure 2 showing the color legend and figure 3 the according slice processing.

Figure 5 shows a block diagram illustrating schematically an exemplary state-transition-cashflow-modelling structure and/or 3 state-transition-cashflow- modelling structure.

Figures 6 to 7 show block diagrams illustrating schematically an exemplary overview of how the Markov Chain modelling structure is used, where figure 6 shows a simple Markov Chain modelling example, and figure 7 a simple Markov Chain calculation for premium cashflow linked to state A. Figure 6 and 7 shows an example of a Markov calculation of the present value for 1 contract. If the example is applied to 1000 customers, then: 1st year, 100*1000 = 100,000 are received. If in 2nd year 20% are deceased, then the active customers are 800 and 90 * 800 = 72,000 are received (instead of 90 * 1000 = 90,000) .

Figure 8 shows block diagrams illustrating schematically exemplary logical architecture of the inventive digital system/platform and method.

Figure 9 shows a block diagram illustrating schematically the data model entities which are used by the calculation service providable by the digital platform apart from the inventive data processing and forecast: (i) Holding all risk-transfer and pricing parameters, (ii) The configuration elements used by the calculation engine to automatically generate quotes, premiums, or refunds parameter values for customers, (iii) Flexible configuration of cashflows attached to states and state transitions, which's based on stochastic processes. However, implementing it in practice for risk-transfer purposes is not known as well as its embedding in the context of the proposed wider solution, (iv) Configuration-based calculations - a feature that allows key calculations to be configured rather than hard-coded, e.g. 'sum-date-years', (v) Benefit slices - To allow maximum flexibility as user needs change, every instance of a policy is made up of multiple benefit slices... the Calc service can calculate on a per slice basis, ensuring accurate and auditable automated pricing parameters generation and outputs.

Figure 10 shows block diagrams illustrating schematically exemplary technical integration structure and pattern for associating 3 ^rd party systems.

Figure 1 1 shows block diagrams illustrating schematically exemplary calculation framework and product framework. The calculation framework comprises

(A) the calculation components of (1 ) Calculation requests are sets of data, which refer to a particular process, product and policy (unless the calculation request is about creating a policy), (2) Calculation rules are sets of rules for calculating values of attributes and attribute properties, and (3) Calculation responses are sets of data, which represent a particular policy; and (B) the calculation entities of (1 ) Calculation entities are sets of attributes and their attribute properties, which are subject to certain calculation rules. The diagram on the right-hand side shows calculation-relevant entity relationships from different domains. Further the product framework defines the full set of allowable configuration options on product/benefit/tariff level, of which subsets can be used by products: (A) The product entity framework defines the full set of allowable configuration options on product level, of which subsets can be used by products. Example: The initial product framework might allow for Premium frequency yearly and monthly and might be enhanced over time to also allow for half-yearly and quarterly.

(B) The benefit entity framework defines the full set of allowable configuration options on benefit level, of which subsets can be used by benefits. Example: The initial benefit framework might allow for Cover type "level" and might be enhanced over time to also allow for "decreasing", and (C) The tariff entity framework defines the full set of allowable configuration options on tariff level, of which subsets can be used by tariffs. Example: The initial tariff framework might allow for Decreasing frequency yearly for the sum assured decreasing sum assured and might be enhanced over time to also allow for monthly. In general calculation-relevant attributes will be located on tariff level to enhance flexibility to allow for different values per policy slice. Example: Premium rate tables are located on tariff level in order to have the ability to always apply the latest Premium rate tables per policy slice.

Figures 12a to 12c show block diagrams illustrating schematically the sequence of steps how a policy with the benefits and the benefits with the slices are being generated. Figures 12a to 12c support the understanding of the description of a process, given by: (A) In order to achieve extensibility (ease of change and extension) and decoupling of the configuration of the calculation from its execution: (i) All calculation elements are, or can be expressed, as a function that receives as argument the current quote period and returns a decimal number. See separate pdf for the full list of functions, (ii) The returned 'BigDecimal' allows to represent any kind of number: double, float, integer... So internally all operations will be done using 'BigDecimal' ...& all defined as a ConfigurationAttribute will be resolved to a CalculationElement. (iii) Example 'special' functions: Present Value; Sum Assured config; (B) The output is a collection of trees of calculation elements: (i) Any output element can be expressed as the composition of some basic calculation elements and its dependent calculation elements. A calculation element can be seen as the root of a tree of all the dependencies it needs to return a decimal number. The leaf nodes of that tree will be always an INPUT or PARAMETER or any other independent calculation element, (ii) ConfigurationAttribute and relationship with CalculationElement, (iii) All ConfigurationAttribute will define a Function<CoverPeriod, BigDecimal>, (iv) Input: The Calculatorinput interface defines the input for our Calculator. We have opted for an implementation of a Typesafe heterogenous container, (v) Output: The Calculatoroutput interface defines the output of our Calculator - to be flexible enough for any kind of Quote model, the output is a mere dictionary of names associated to Collection<BigDecimal>. (vi) Validation is also a function: CalculationElementValidator: The CalculationElementValidator interface it is also a function, but in this case is a Predicate<CoverPeriod, BigDecimal>. Two implementations are provided, but there is not implementation for the parsing from ConfigurationAttribute to simplify the PoC (the implementation of them could be easily added in the CalculationFactory because Calculation offers them in its API) .

Figure 13 shows block diagrams illustrating schematically an exemplary calculation and generation process, respectively. They support the understanding of the description of a process, given by (note that not all in this description is part of the diagram): (i) After all CalculationElement are created from their Configuration. getAIIAttributes() counterparts, the created Calculation is passed to the Calculator, (ii) Quote calculation' included in above diagram, (iii) Slice management/built quote is also covered in figure 13. In addition, the digital system can e.g. comprise (a) Stream parallelization in Calculator, (b) Returning already built CalculationElement in CalculationFactory + More implementations of CalculationElement, (c) coverTermYears..., not having to calculate all the elements again (d) Enhance calculation request configuration flexibility by atomizing the calculation request into micro calculation requests on quote, benefit and slice level, which are pre-processed and run in dynamic order, so that MTAs can be defined exclusively by pre-processing rules (and all of them use the same core-processing rules) : (i) pre-processing i.e. creating the micro calc requests on quote, benefit and slice level and populate them, (ii) core-processing, i.e. executing the micro calc requests by using the core calculation, which always stays the same.

Figure 14 shows block diagrams illustrating schematically an exemplary technical accounting (TA) structure by a process flow diagram of the solution (note that not all in this description is part of the diagram) . The goal of the technical accounting domain is to provide source accounting data in order to support ledger systems (LFD). The solution allows an event-based policy admin system to integrate with traditional RDBMS (Relational DataBase Management Systems) LFD platforms. Further it allows to provide outputs policy-level accounting data via both DLE & Control Report, allowing data to be cross-checked directly with the source system.

Figure 15 shows block diagrams illustrating schematically an exemplary technical accounting (TA) structure (note that not all in this description is part of the diagram). The embodiment variant uses (i) DDD (Domain Driven Design), (ii) CQRS (Command Query Responsibility Segregation) (iii) ES (Event Sourcing) . It aggregates with ES: (a) Accounting Plan (b) Remittances, (c) Arrears. It aggregates without ES: (a) Quote. ES + CQRS is implemented with a lightweight framework. ES Events are internal, and are NOT events in Kafka topics, but events in an internal event store. Only belongs to the Technical Accounting service. The system uses Hexagonal architecture, the project can be split into modules by concerns. The Domain only knows about the interface of the persistent store but has no idea about the underlying storage / implementation. Figure 16 shows block diagrams illustrating schematically an exemplary policy lifecycle processes (note that not all in this description is part of the diagram) . The policy structure and service hold policy-related data and drives the policy lifecycle, listening to all other services and making sense of the policy /customer at any given point in time. It includes the third party service (i.e. customer/company data and their roles against individual policies). The embodiment variant e.g., comprises (i) The definable rules, behaviour and interaction of the 'Product/Calc' and 'Policy' services, as well as the operations and validations required to get take a policy into force, plus the associated benefits and riders - all of which is driven by configuration, (ii) Benefit slices - To allow maximum flexibility as customers' needs change, every instance of a policy is made up of multiple benefit slices. While slices exist in other systems, the inventive use and interact with them across the system to generate policy changes (Mid Term Adjustments) is not known in the prior art.

Figure 17 shows block diagrams illustrating schematically an exemplary policy structure. Design principles of the refactored policy structure and service design can e.g. comprise (i) Isolate application and domain logic, (ii) Domain-driven-design, (iii) Improved REST API for policy management to follow the business processes, (iv) Evolving of the event schemas to simplify messaging.

Figure 18 shows a block diagram illustrating schematically an exemplary digital system and platform 1 for mortality probability parameter value propagation 121 and for dynamic and automated impact-cover pricing 122 by processing a plurality of individual mortality-related measuring parameters 123 associated with a portfolio 124 of loss-covers 1241 of risk-exposed individuals 2. Each risk-transfer 1241 held associated with the portfolio 124 is set by risk-transfer parameters 1241 1 of a loss-covers 1241 as risktransfer policy defining parameter-based the individual risk-transfer 1241 , wherein a combination of a Markov Chain modelling structure 1 1 with configurable elements are applied at least comprising states 1 14 and/or state transitions 1 13 and/or cash-flows 1 12 specific to the risk-transfer 1241 , and wherein a stochastic Markov data processing is applied to the Markov chain structure 1 1 over a sequence of possible events 25/25i in which the probability value of each event 25/25i depends only on the state 1 14 attained in the previous event 25. Figure 19 shows a block diagram illustrating schematically the triggering of an automated digital twin structure adaption steered by output signal 143 generated by digital individual measuring engine 3 based on measured parameter values of wearables/loT sensory 15, and on the other side the feedback of monitored and/or detected by thresholding of the digital twin parameters to the simulation engine 12 and/or the Markov Chain structure 1 1 .

Detailed Description of the Preferred Embodiments

Figures 1 -17 schematically illustrate an architecture for a possible implementation of an embodiment of the inventive digital platform and digital system 1 for mortality probability measuring parameter value propagation, i.e. measured life risk parameter propagation, and for dynamic and automated risk-transfer pricing by processing a plurality of individual risk-related parameters associated with a portfolio of risk-transfers of risk-exposed individuals. The pricing is dynamically performed and also the dynamic generation of the risk-transfer premiums for an individual policy using a combination of a Markov Chain modelling structure and configurable elements at least comprising states and/or state transitions and/or cash-flows specific to the product it instantiates. Each risk-transfer held associated with the portfolio is set by risk-transfer parameters of a risk-transfer policy defining the individual risk-transfer. A combination of a Markov Chain modelling structure with configurable elements are applied at least comprising states and/or state transitions and/or cash-flows specific to the risk-transfer. A stochastic Markov data processing is applied to the Markov chain structure over a sequence of possible events in which the probability value of each event depends only on the state attained in the previous event.

The digital system comprises a calculation engine using deterministic transition and interest rates of a discrete process to conduct calculations per policy, wherein parts of the process are used with stochastic transition and interest rates of a discrete process for conducting simulations on portfolio level. In particular, the calculation engine comprises a data structure for capturing and/or storing deterministic transition and interest rate parameter values of a state discrete process to conduct data processing per policy, wherein parts of the process are used with stochastic transition and interest rates of a discrete process for conducting simulations on portfolio level,

The digital system comprises a user interface to the Markov chain structure, wherein for the stochastic Markov data processing, interest parameter values and transition rate parameter values are user-specific and flexible configurable and/or selectable from an associated digital library, and wherein within the stochastic Markov data processing setup and stochastic transition rates, these underlying rates are processed and modelled by the finite-state Markov chain structure. The digital library can e.g. be accessible by a plurality of users, wherein the transition functions and/or the interest parameter values and/or the transition rate parameter values of a user are accessible by another user via the data interface and applicable to the other user's finite-state Markov chain structure. The transition functions and/or the interest parameter values and/or the transition rate parameter values of a first user can e.g. only be accessible upon request of a second user and/or upon approval or enablement by the first user to the second user.

For the data processing by the finite-state Markov chain structure, one or more transition functions are configurable via the data interface and/or selectable from the digital library, the transition functions linking at least two states within the Markov chain structure wherein all states of the Markov chain structure are linked to an antecedent and a successive state providing the data processing over the whole configurable Markov chain structure.

As an embodiment variant, dependences between the rate parameters can e.g. be applicable within the finite-state Markov chain structure. The finite-state Markov chain structure can e.g. further comprise elements providing flexible configuration for the use of combined model structures for stochastic interest parameter values and/or mortality rate parameter values. The stochastic Markov data processing of the interest parameters and mortality rate parameters can be configured by affine data process structures, wherein the finite-state Markov chain structure becoming a traceable model structure during propagation of the parameter values to a defined future time window. For the flexible configuration, the digital system can e.g. comprise adaptable calculation configuration files and/or trees processable by the calculation engine of the digital system. The Markov chain structure can e.g. be realized as a continuous time Markov chain structure with a finite or countable infinite state space providing a stochastic process for parameter propagation. Further, the Markov chain structure can e.g. be provided by the digital system only for time- homogeneous Markov chain processes, where all probability values providing a measure for a life risk within a future time window are generated based on the Markov property.

The system can e.g. comprise a slice generator, wherein data processing is sliced by taking place on a slice level and aggregated on a state of benefit and quote level, wherein a slice is created whenever a policy benefit is subject to an unscheduled sum assured increase, and wherein for each newly applied risk-transfer each benefit starts with one slice, each slice being related to the specific risk-transfer version, cover and tariff at that a respective point in time, the slice is generated for. As such, the calculations can e.g. take place on a slice level and are aggregated on benefit and quote level, wherein a slice is created whenever a policy benefit is subject to an unscheduled sum assured increase, and wherein for new business each benefit starts with one slice, each slice being related to the specific product version, cover, and tariff relevant at that the respective point in time the slice is created for.

The present invention allows insurance premiums to be generated and assessed dynamically for an individual policy, using a combination of the established Markov Chain model and configurable elements (States, State transitions, Cashflows) specific to the product it instantiates. The system comprises a calculation engine using deterministic transition and interest rates of a discrete process to conduct calculations per policy. The system can e.g. also use parts of this with stochastic transition and interest rates of a discrete process can thus support conducting simulations on portfolio level.

The inventive data processing and/or calculations take place on slice level and are aggregated on benefit and quote level (as shown in figure 1 ) . A slice is created whenever a policy benefit is subject to an unscheduled sum assured increase, i.e. for new business each benefit starts with one slice. Each slice relates to the specific product version, i.e. risk-transfer structure version, cover and tariff that is relevant at that the respective point in time, the slice is created for. While Markov Chain modelling structures are used widely in automated insurance and risk-transfer technology and other industries, its combination with configuration elements allows a level of flexibility and dynamism that is novel and not known in the prior art.

Figures 2 to 4 show block diagrams illustrating schematically an exemplary overview of slices with figure 2 showing the color legend and figure 3 the according slice processing.

The present invention also technically allows to model cashflows and link them to states or state transitions according to the respective tariff - this allows to generate and/or calculate the cashflows' present value at each discrete point in time. This dynamic is visualized in figure 5. Figure 5 shows a block diagram illustrating schematically an exemplary state-transition-cashflow-modelling structure and/or 3 state-transition-cashflow-modelling structure. Figures 6 to 7 show block diagrams illustrating schematically an exemplary overview of how the Markov Chain modelling structure is used, where figure 6 shows a simple Markov Chain modelling example, and figure 7 a simple Markov Chain calculation for premium cashflow linked to state A. Figure 6 and 7 shows an example of a Markov calculation of the present value for 1 contract. If the example is applied to 1000 customers, then: 1st year, 100*1000 = 100,000 are received. If in 2nd year 20% are deceased, then the active customers are 800 and 90 * 800 = 72,000 are received (instead of 90 * 1000 = 90,000) .

A mathematical representation of the present value formula for a (sub-)set of modelled cashflows on slice level is as follows:

PV _t ^s (CF) ... Present value of a set CF of cashflow amount vectors for state S at calculation period t

C ... Cash flow amount for state S at calculation period t v ... Discount rate at time t for the time from calculation period t to calculation period t + k q;^ ^T ... Transition rate for state S to state T at calculation period t

C _t ^s ^ ^T ... Cashflow amount for state S to state T at calculation period t It is to be noted that the implementation of the calculation engine can e.g. use deterministic transition and interest rates of a discrete process to conduct calculations on policy level. However, the present invention can also be realized using parts of this with stochastic support conducting the simulation processing on portfolio level.

List of reference signs Digital Platform and System

10 Data Store

101 Data Structure

101 1 Modular Digital Transition Functions

1012 Markov Chain Data Elements

1 1 Markov Chain Modelling Structure/Predictor/Simulator

1 1 1 Applied Transition Functions

1 12 Applied Interest Parameter Values

1 13 Applied Transition Rate Parameter Values

1 14 Markov Chain States

1 15 Output parameter values of the flexible Markov Chain Structure

12 Simulation engine for automated process / Stochastic Markov data processing

121 Probability parameter value propagation process

122 Automated pricing process

123 Individual mortality-related measuring parameters based on measuring parameters associated with individual 2/2i/2ii

124 Data structure container / Portfolio

1241 Data structures capturing loss-cover/risk-transfer parameters of a risk-transfer (Policy/Risk-Transfer)

1241 1 Risk-transfer parameters

125 Digital library

13 Data transfer interface

14 Output signal generator

140 Electronic output signaling

141 Trigger signal triggering automated underwriting process

141 1 Automated electronic underwriting system

142 Steering and thresholding signal for automated portfolio management and management adaption

1421 Automated digital portfolio management system electronically steered by the signaling of the Markov Chain Predictor 1 1

143 Triggered automated digital twin adaption steered by output signal generated by digital individual measuring engine 3 based on measured parameter values of wearables/loT sensory 15

144 Digital actionable offers

145 Automatedly generated new digital risk-transfer policies

15 Wearables/loT Sensory (input devices and sensors)

151 Bodily sensory devices

152 Environmental sensory devices

16 Data transmission network Real-world Risk-exposed Individual

21 Physical Body Structure

22 Intangible Body Conditions (Psychological)

23 Subsystems of the Real-world Individual

231 , 232, 233 23i Subsystems 1 i

24 Environment/Ecosystem of Individual

241 , 242, 243 24i Subsystems 1 i

25 Physical real-world event impacting physically individual by causing a measurable physical damage/body injury or loss to the individual

251 Illness

252 Accident

253 Natural catastrophe (earthquake/storm/hurricane/flood etc.) Digital Individual Measuring Engine

31 Digital Intelligence Layer

31 1 Machine Learning

312 Neural Network

32 Body Parameters of the Real-World Individual

33 Status Parameters of Real-World Individual

331 Physiological (Body) Status Parameters

332 Psychological Status Parameters

333 Habits/Behavioral Status Parameters (Nutrition/Sport etc.)

333 Environmental Status Parameters (Living Condition etc.)

34 Data Structures Representing States of Each of the Plurality of Subsystems of the Real-World Individual

35 Digital Peril and/or Life-risk-event Robot

351 Simulation

352 Synchronization 353 Twin Linking: Sensory/Measuring/Data Acquisition - Wearables/loT

Sensory 15 Digital Ecosystem Replica Layer

361 , 362, 363 36i Virtual Subsystems of Virtual Representation of Ecosystem Digital Object/Element Layer of the Individual

371 Simulation

372 Synchronization

373 Linking: Sensory/Measuring/Data Acquisition - Wearables/loT Sensory 15 Digital Individual Replica Layer

381 , 382, 383 38i Virtual Subsystems of Real-World Individual