Asset option information generation method and device, electronic equipment and storage medium

Abstract

The invention discloses an asset option information generation method and device, electronic equipment and a storage medium, and relates to the field of quantum computing, and the method comprises the steps: obtaining a historical actual price path of an underlying asset in an asset market; and inputting the historical actual price path to a preset quantum neural network model, and outputting option information of the underlying assets under the historical actual price path through the preset quantum neural network model, the option information including option optimal pricing data and a hedging strategy model. The objective of the invention is to solve the technical problem that option pricing of underlying assets and prediction accuracy of a hedging strategy model are low.

Description

Technical Field

[0001] The present application relates to the field of quantum computing, and in particular to a method, device, electronic device, and storage medium for generating option information of an asset. Background Art

[0002] Option pricing and hedging strategy models are key issues in the field of financial derivatives. Options allow holders to buy or sell an underlying asset at a specific price in the future. Hedging strategy models can help investors lock in profits, limit losses, and reduce portfolio volatility, ultimately protecting their investments and maximizing profits through proper risk management.

[0003] With the development of machine learning, more and more research is exploring the use of these technologies to improve option and hedging strategy models. Currently, traditional machine learning is often used to predict the underlying asset's option pricing and hedging strategy models. However, when faced with high market volatility and nonlinear characteristics, traditional machine learning methods have difficulty fully capturing the complex market changes. As a result, traditional machine learning methods have low prediction accuracy, making it difficult to effectively manage the risk of the underlying assets, and the prediction of the underlying asset's option pricing and hedging strategy models is not accurate enough.

[0004] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a method, device, electronic device and storage medium for generating option information of an asset, aiming to solve the technical problem of low prediction accuracy of option pricing and hedging strategy models of underlying assets.

[0006] To achieve the above objectives, the present application provides a method for generating asset option information, the method comprising:

[0007] Obtain the historical actual price path of the underlying asset in the asset market;

[0008] The historical actual price path is input into a preset quantum neural network model, and the preset quantum neural network model outputs option information of the underlying asset under the historical actual price path, wherein the option information includes option optimal pricing data and a hedging strategy model.

[0009] In one embodiment, the method comprises:

[0010] Obtaining historical market data for the underlying asset;

[0011] simulating a price path of the underlying asset based on the historical market data;

[0012] An initial quantum neural network model is obtained, and the initial quantum neural network model is trained according to the price path to obtain a trained preset quantum neural network model.

[0013] In one embodiment, the step of simulating the price path of the underlying asset based on the historical market data includes:

[0014] Get the preset price simulation model and the preset time step;

[0015] Obtaining price simulation parameters required by the preset price simulation model from the historical market data;

[0016] Configuring the price simulation parameters and the preset time step in the preset price simulation model;

[0017] Based on a preset price simulation model and a preset random numerical method, the price path of the underlying asset is simulated.

[0018] In one embodiment, the step of simulating the price path of the underlying asset based on a preset price simulation model and a preset random value method includes:

[0019] Generate a random path for the underlying asset based on a preset random value method;

[0020] The preset price simulation model is used to generate the asset price of the underlying asset at each preset time step in the random path, thereby obtaining the price path of the underlying asset.

[0021] In one embodiment, the step of obtaining an initial quantum neural network model includes:

[0022] Get the hedge type of the underlying asset;

[0023] Constructing an optimal pricing mathematical model for the underlying asset based on the hedge type, a preset objective function, and the transaction risk constraints of the underlying asset;

[0024] A quantum neural network structure is obtained, and when the quantum neural network structure is in a preset approximation state, an initial quantum neural network model of the optimal pricing mathematical model is determined based on the quantum neural network.

[0025] In one embodiment, the step of training the initial quantum neural network model according to the price path to obtain a trained preset quantum neural network model includes:

[0026] Determining, based on the initial quantum neural network model, a training hedging strategy model corresponding to each preset time step in the price path, and determining a training option pricing for the underlying asset;

[0027] Calculating, based on the optimal pricing mathematical model, a training loss value corresponding to each of the training hedging strategy models and the training option pricing;

[0028] The initial quantum neural network model is updated based on the training loss value, and when the training loss value is less than a preset loss threshold, the initial quantum neural network model is used as a preset quantum neural network model.

[0029] In one embodiment, after the step of inputting the historical actual price path into a preset quantum neural network model and outputting, through the preset quantum neural network model, option information of the underlying asset under the historical actual price path, wherein the option information includes option optimal pricing data and a hedging strategy model, the method further includes:

[0030] Based on the optimal option pricing data and the hedging strategy model, performing a hedging operation on the underlying asset to obtain a hedging result;

[0031] The preset quantum neural network model is optimized based on the hedging result.

[0032] In addition, to achieve the above-mentioned purpose, the present application also provides an asset option information generation device, which includes:

[0033] The acquisition module is used to obtain the historical actual price path of the underlying asset in the asset market;

[0034] An analysis module is configured to input the historical actual price path into a preset quantum neural network model, and output, through the preset quantum neural network model, option information of the underlying asset under the historical actual price path, wherein the option information includes optimal option pricing data and a hedging strategy model.

[0035] In addition, to achieve the above-mentioned purpose, the present application also provides an electronic device, which includes: a memory, a processor, and a program of the method for generating option information of the asset stored on the memory and runnable on the processor. When the program of the method for generating option information of the asset is executed by the processor, the steps of the method for generating option information of the asset as described above can be implemented.

[0036] In addition, to achieve the above-mentioned purpose, the present application also provides a computer-readable storage medium, on which is stored a program for implementing the method for generating asset option information. When the program for the method for generating asset option information is executed by a processor, the steps of the method for generating asset option information as described above are implemented.

[0037] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for generating asset option information.

[0038] One or more technical solutions proposed in this application have at least the following technical effects: This application obtains the historical actual price path of the underlying asset in the asset market, inputs the historical actual price path into a preset quantum neural network model, and outputs option information for the underlying asset under the historical actual price path through the preset quantum neural network model, wherein the option information includes optimal option pricing data and a hedging strategy model. Because this embodiment can determine the optimal option pricing data and hedging strategy model corresponding to the underlying asset based on the preset quantum neural network model, and the quantum neural network model is based on quantum computing, the parallelism of quantum computing and the superposition of quantum states corresponding to quantum computing enable the quantum neural network to more effectively capture the nonlinear characteristics of the underlying asset in the asset market based on the historical actual price path, thereby facilitating improved accuracy in determining the option pricing and hedging strategy model for the underlying asset. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application, and together with the description, serve to explain the principles of the present application.

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0041] Figure 1 This is a flow chart of an embodiment of a method for generating option information for assets of the present application;

[0042] Figure 2 A flowchart illustrating an example of a method for generating option information for assets in this application;

[0043] Figure 3 A flowchart of another example of a method for generating option information for assets in this application;

[0044] Figure 4This is a schematic diagram of a module flow chart of an example of a method for generating option information of an asset in this application;

[0045] Figure 5 A module flow chart of another example of a method for generating option information for assets in this application;

[0046] Figure 6 This is a structural diagram of an embodiment of a device for generating option information for assets of the present application;

[0047] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the method for generating asset option information in the embodiment of the present application.

[0048] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0049] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0050] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0051] Option pricing and hedging strategy models are key issues in the field of financial derivatives. Accurate pricing and effective and rational hedging strategy models can help investors reduce or even theoretically eliminate the risks associated with adverse price fluctuations. Option pricing reflects the market's consensus expectations regarding future uncertainties. An effective option pricing model that reasonably reflects the future price volatility and time value decay of the underlying asset is a key indicator of financial market efficiency. Through correct option pricing, market participants can better discover prices, manage risks, and achieve optimal asset allocation. Options, as derivatives, allow holders to buy or sell the underlying asset at a specific price in the future, providing investors with an effective risk management tool. The use of hedging strategy models can help investors lock in profits, limit losses, and reduce portfolio volatility, thereby protecting their investments and maximizing profits through appropriate risk management. Option pricing and hedging strategy models are fundamental to financial innovation and the design of new financial products. The design and evaluation of many complex financial derivatives and structured products rely on the understanding and pricing of options. Effective option pricing and hedging strategy models encourage more market participants to enter the market, increasing market depth and liquidity. Option pricing and hedging strategy models not only help investors effectively manage risks and allocate assets, but also promote the development and innovation of financial markets and improve the overall efficiency and functionality of the market.

[0052] Traditional mathematical models used for option pricing include the Black-Scholes model (BS model). This model, proposed by Fischer Black, Myron Scholes, and Robert Merton in 1973, is used to price European options. The BS model assumes that the price of the underlying asset follows a geometric Brownian motion and takes into account factors such as the risk-free rate, the volatility of the underlying asset, the strike price, and the remaining time to maturity. The risk-free rate is the interest rate that investors can earn without risk. In finance, the risk-free rate is often used to measure the discount rate for assets and to calculate the present value of options. Volatility is a measure of the price fluctuations of the underlying asset, reflecting the degree of uncertainty in the underlying asset's price.

[0053] Despite its historical success, traditional models of option pricing and hedging strategies have limitations when dealing with complex market environments and financial instruments. For example, traditional models of option pricing and hedging strategies struggle to adapt to complex real-world scenarios. These models, which are based on linear results discussed and calculated under ideal conditions, are limited in their ability to handle high-dimensional and complex problems. This limits their application and effectiveness in pricing complex financial derivatives.

[0054] Traditional machine learning-based option pricing methods, while offering some improvements over traditional mathematical models for handling some complex problems, remain computationally prohibitively expensive when faced with extremely complex financial market data. Training parameters are often too large and training time is prohibitively long. While machine learning methods can improve forecasting accuracy by learning from historical data, they are limited in their accuracy for predicting option pricing and hedging strategies when faced with the high volatility and nonlinearity of the market. Traditional machine learning models struggle to fully capture the complex dynamics of financial markets and the complex nonlinear relationship between option prices and their underlying asset prices.

[0055] To this end, the embodiment of the present application proposes the use of quantum neural networks to determine the corresponding option prices and hedging strategy models. Quantum neural networks are the product of the combination of quantum computing and neural networks. Using quantum bits for information processing can improve computing efficiency and processing power. Quantum computing can handle complex optimization problems that are difficult for traditional computers to solve. In addition, the introduction of quantum neural networks can reduce the number of neural network parameters, thereby improving the computational efficiency of option pricing. The parallel processing capabilities of quantum computing can significantly shorten the time required for quantum neural network model training and prediction.

[0056] Furthermore, due to the parallelism of quantum computing and the superposition of quantum states corresponding to quantum computing, quantum neural networks can more effectively capture the nonlinear characteristics of underlying assets in asset markets based on historical actual price paths, thereby improving the accuracy of option pricing and hedging strategy models for determining underlying assets. Furthermore, they can better process high-dimensional and complex financial data.

[0057] The embodiment of the present application provides a method for generating asset option information, referring to Figure 1 , Figure 1 This is a flow chart of an embodiment of a method for generating asset option information of the present application. In this embodiment, the method for generating asset option information includes steps S10 to S20:

[0058] Step S10, obtaining the historical actual price path of the underlying asset in the asset market;

[0059] It should be noted that the underlying asset is an asset that can be bought and sold. The underlying asset can be a stock, stock index, currency, commodity, bond, or other financial instrument, etc., and this embodiment does not specifically limit this. The asset market refers to the place where the underlying asset is traded. The asset market allows investors to buy and sell the underlying asset.

[0060] The historical actual price path is a sequence of price changes of the underlying asset over time. It records the price fluctuations of the underlying asset over time. This path can be used to determine the future price fluctuations of the underlying asset, facilitating the subsequent determination of optimal option pricing data and hedging strategy models. The optimal option pricing data can be the current optimal price or the optimal price at the last moment in the historical actual price path. A hedging strategy model is characterized as a strategy that reduces the risk of option trading by buying or selling the underlying asset or its derivatives.

[0061] Illustratively, market data of the underlying asset in the asset market can be obtained, such as historical market data, etc. The asset price, volatility, and risk-free interest rate of the underlying asset can be obtained from the historical market data. The asset price, volatility, and risk-free interest rate of the underlying asset in the historical market data can be cleansed and normalized, and the historical actual price path of the underlying asset can be determined based on the normalized asset price, volatility, and risk-free interest rate.

[0062] Volatility is a statistical indicator that measures the magnitude of price fluctuations in an underlying asset. It can be used to reflect the severity of price fluctuations. Higher volatility indicates more volatile price fluctuations and greater uncertainty about future prices. Lower volatility indicates relatively stable prices and less uncertainty about future prices. The risk-free rate is the theoretical rate of return on an investment that is free of default risk.

[0063] In step S20, the historical actual price path is input into a preset quantum neural network model, and the preset quantum neural network model outputs option information of the underlying asset under the historical actual price path, wherein the option information includes the option optimal pricing data and the hedging strategy model.

[0064] It should be noted that the preset quantum neural network model is used to determine the optimal pricing data for options on the underlying asset and the hedging strategy model. The preset quantum neural network model can be pre-trained. Options give the option holder the right, but not the obligation, to buy or sell the underlying asset at a specific price (strike price) at a specific time in the future. The option buyer pays a fee (premium), while the seller collects this fee and assumes the corresponding obligations. Options can be divided into various types, such as European options, American options, and barrier options.

[0065] Option information may include optimal option pricing data and hedging strategy models. The optimal option pricing data is the most ideal price of the option corresponding to the underlying asset. The optimal option pricing data can be the optimal pricing at the current moment or the optimal pricing at the last moment corresponding to the actual historical price path.

[0066] A hedging strategy model is characterized as a strategy model that reduces the risk of option trading by buying or selling the underlying asset or its derivatives. For example, the hedging strategy model may include the holdings of the underlying asset. The hedging strategy model can be represented as an algorithm for hedging or the steps for executing hedging. This embodiment does not specifically limit the representation of the hedging strategy model. Hedging strategy models can be divided into multiple types. For example, hedging strategy models may include delta hedging, delta neutral hedging, and gamma hedging. Delta hedging adjusts the holdings of the underlying asset to offset or reduce the sensitivity of options or other derivatives to changes in the underlying asset's price. Delta neutral hedging reduces the total delta value of the portfolio corresponding to the underlying asset to zero. That is, regardless of whether the underlying asset's price rises or falls, the total value of the portfolio remains unchanged. The core concept of this hedging strategy model is to adjust the holdings of the underlying asset so that the portfolio's sensitivity to changes in the underlying asset's price (i.e., the delta value) approaches zero. The delta value is an indicator that measures the value of an option relative to changes in the underlying asset's price. The purpose of gamma hedging is to reduce or eliminate the risk brought by the gamma value by adjusting the option positions in the portfolio. The gamma value measures the rate at which the delta value changes with the price of the underlying asset.

[0067] Different underlying assets may require different types of options, and may require different types of hedging strategy models. Different option types may also require different preset quantum neural networks, and different hedging strategy models may require different preset quantum neural network models.

[0068] For example, the historical actual price path can be input into a preset quantum neural network model with the same type of hedging strategy model and option type as those required for the underlying asset. The preset quantum neural network model can perform risk analysis on the historical actual price path to obtain the optimal option pricing data of the underlying asset and the corresponding hedging strategy model.

[0069] This embodiment of the present application obtains the historical actual price path of the underlying asset in the asset market, inputs the historical actual price path into a preset quantum neural network model, and outputs option information for the underlying asset under the historical actual price path through the preset quantum neural network model, where the option information includes optimal option pricing data and a hedging strategy model. Because this embodiment can determine the optimal option pricing data and hedging strategy model corresponding to the underlying asset based on the preset quantum neural network model, and the quantum neural network model is based on quantum computing, the parallelism of quantum computing and the superposition of quantum states corresponding to quantum computing enable the quantum neural network to more effectively capture the nonlinear characteristics of the underlying asset in the asset market based on the historical actual price path, thereby facilitating improved accuracy in determining the option pricing and hedging strategy model for the underlying asset.

[0070] In a feasible embodiment, the method for generating asset option information further includes steps A10 to A40:

[0071] Step A10, obtaining historical market data of the underlying asset;

[0072] It should be noted that historical market data refers to market data of the underlying asset in the asset market. Historical market data may include price drift rate, volatility, and the asset price of the underlying asset. Volatility may include dynamic volatility, etc. This embodiment does not specifically limit this. Different underlying assets may also have different historical market data. For example, some underlying assets may not have parameters such as price drift rate.

[0073] Step A20, simulating the price path of the underlying asset based on historical market data;

[0074] It should be noted that multiple price paths for the underlying asset can be simulated based on historical market data, and each price path may not be the actual price path of the underlying asset. Each simulated price path can be used as a sample for training the initial quantum neural network model. For example, multiple price paths can be simulated based on historical market data, each with a different price path.

[0075] In a feasible embodiment, step A20 further includes steps A21 to A24:

[0076] Step A21, obtaining a preset price simulation model and a preset time step;

[0077] It should be noted that the preset price simulation model is used to simulate the price path of the underlying asset. Different underlying assets may correspond to different preset price simulation models. The preset price simulation model can be determined based on actual circumstances. For example, the preset price simulation model can be determined based on market characteristics reflected by the underlying asset's historical market data. This embodiment does not specifically limit this. Market characteristics may include, for example, the nature of price fluctuations and the trading volume of the underlying asset. For example, the preset price simulation model may be a geometric Brownian motion model, a stochastic volatility model, a jump-diffusion model, etc.

[0078] The preset time step can be determined based on actual conditions. For example, the preset time step can be determined based on simulation accuracy and computational cost. The longer the preset time step, the lower the simulation accuracy and the lower the computational cost. The shorter the preset time step, the higher the simulation accuracy and the higher the computational cost.

[0079] Step A22, obtaining price simulation parameters required by a preset price simulation model from historical market data;

[0080] It should be noted that different preset price simulation models may not necessarily correspond to the same price simulation parameters. Price simulation parameters can be determined based on the preset price simulation model. For example, the BS model can directly use the volatility and interest rate of the underlying asset to simulate the price path, or it can calculate the volatility and interest rate through historical asset prices to simulate the price path through volatility and interest rate. Other models, such as the stochastic volatility model, may require more parameters for price path simulation, such as implied volatility, price jump frequency, etc. For example, the parameter names of the required price simulation parameters can be determined through the preset price simulation model, and the parameter values ​​of the price simulation parameters can be obtained from historical market data based on the obtained parameter names. This is not specifically limited in this embodiment.

[0081] Step A23, configuring the price simulation parameters and the preset time step in the preset price simulation model;

[0082] Step A24: Simulate the price path of the underlying asset based on a preset price simulation model and a preset random value method.

[0083] It should be noted that the preset random number method can be an Euler numerical method, a Monte Carlo simulation method, etc., and the preset random number method can be used to generate multiple random numbers. Price simulation parameters and preset time compensation can be configured in the preset price simulation model to facilitate subsequent simulation of the price path of the underlying asset.

[0084] For example, a preset price simulation model and a preset time step are obtained, and the preset time step and price simulation parameters obtained from historical market data are configured in the preset price simulation model. For example, the price simulation parameters may be volatility, interest rate, etc. Based on the preset price simulation model and the preset random value method, the price path of the underlying asset is simulated, and multiple price paths of the underlying asset may be simulated.

[0085] This embodiment allows the price simulation parameters and preset time step to be configured in a preset price simulation model, thereby facilitating the subsequent simulation of multiple price paths and, in turn, the subsequent training of the initial quantum neural network model. Furthermore, because the preset time step can be configured based on actual conditions, the flexibility of price path simulation is also increased.

[0086] In a feasible embodiment, step A24 further includes steps A241 and A242:

[0087] Step A241: Generate a random path for the underlying asset based on a preset random value method;

[0088] In step A242, the asset price of the underlying asset at each preset time step in the random path is generated by a preset price simulation model to obtain the price path of the underlying asset.

[0089] It should be noted that the random path is a path of random numbers generated by a preset random numerical method. The random path may include multiple random numbers, and the preset random numerical method may generate multiple random paths. The number of random numbers in each random path may be the same, and the random numbers in the same random path may be different. The number of random numbers in the random path may be determined by a preset time step and a preset total path duration. For example, the number of random numbers in the random path may be the ratio of the preset path total duration to the preset time step. The preset path total duration may be customizable, and this embodiment does not specifically limit this. The preset path total duration may be used to describe the path length of the price path, or may be used to describe the path length of the random path. The path length is characterized by duration.

[0090] The path length of the random path is also the total time length of the preset path. The preset random value method can generate random numbers corresponding to a preset number of preset time steps, and splice the random numbers corresponding to each preset time step to obtain a random path.

[0091] The preset price simulation model can generate the price of the underlying asset at each preset time step in a random path, thereby obtaining a simulated price path for the underlying asset. For example, the preset price simulation model can calculate the asset price at each preset time step based on the random number at each preset time step in the random path, and then concatenate the asset prices at each preset time step to obtain the simulated price path for the underlying asset. Each random path has its own corresponding price path.

[0092] Since this embodiment can generate a random path for the underlying asset based on a random numerical method and can determine multiple random paths, it is convenient to subsequently determine the price path corresponding to the random path. Moreover, since the random path is actually obtained by splicing together various random numbers, it is convenient to generate more random paths, thereby facilitating the simulation of more price paths, thereby facilitating the enrichment of the training samples for training the initial quantum neural network, improving the training accuracy of the initial quantum neural network, and reducing the cost and complexity of obtaining training samples.

[0093] To better understand this embodiment, please refer to the following example to briefly illustrate the price path of the simulated underlying asset in this embodiment. For example, in this example, the preset price simulation model can be a geometric Brownian motion model, the preset numerical method can be an Euler numerical method, and the geometric Brownian motion model can be expressed as:

[0094] dS t =μS t dt+σS t dW t

[0095] Among them, S t is the price of the underlying asset at time t, μ is the expected rate of return of the underlying asset. Since the expected rate of return is difficult to estimate accurately, the risk-free rate r can generally be used instead of the expected rate of return. σ is the volatility of the underlying asset, and dW t is a Wiener process, a stochastic process used to describe the randomness of underlying asset price fluctuations. dt is the time increment, which can be a preset time step. In the geometric Brownian motion model, dSt represents the change in the underlying asset price St within a small time increment dt. Both volatility σ and the risk-free rate r can be price simulation parameters required by the geometric Brownian motion model. Further, using the Euler numerical method, the expression of the above geometric Brownian motion model is converted to discrete-time form, which can be expressed as:

[0096]

[0097] Where Δt is the preset time step, z is a random number, which can be a random variable drawn from a standard normal distribution to simulate the randomness of price changes, and St+Δt is the price of the underlying asset at the next time step t+Δt, exp represents the natural exponential function, and N(0,1) represents the standard normal distribution. The above formula can be used to simulate the price path by first simulating the asset price at each preset time step, then concatenating the asset prices at each preset time step to obtain the price path.

[0098] Step A30: Obtain an initial quantum neural network model, and train the initial quantum neural network model according to the price path to obtain a trained preset quantum neural network model.

[0099] It should be noted that the initial quantum neural network model is a model to be trained, and the initial quantum neural network model can be determined based on the type of options required for the underlying asset and the strategy type of the hedging strategy model.

[0100] For example, the initial quantum neural network model can be iteratively trained based on various price paths until a fully trained preset quantum neural network model is obtained. This facilitates the subsequent use of the preset quantum neural network model to predict the optimal option price and hedging strategy model for the underlying asset.

[0101] This embodiment utilizes the parallel processing capability of quantum computing by training the initial quantum neural network model, thereby reducing the time required to train the initial quantum neural network model.

[0102] In a feasible embodiment, step A30 further includes steps A31 to A33:

[0103] Step A31, obtaining the hedging type of the underlying asset;

[0104] Step A32: constructing an optimal pricing mathematical model for the underlying asset based on the hedging type, the preset objective function, and the transaction risk constraints of the underlying asset;

[0105] It should be noted that the hedging type of the underlying asset can be determined based on the type of hedging strategy model required for the underlying asset. The type of hedging strategy model required for the underlying asset can be determined based on actual circumstances, and this embodiment does not specifically limit this. The preset objective function can be a convex risk measurement function, and the transaction risk constraint conditions include preset trading restrictions and preset maximum risk limits for the underlying asset. For example, the maximum risk limit can be a preset risk threshold, and the investment risk corresponding to the underlying asset must be less than or equal to the preset risk threshold; trading restrictions can include restrictions on the purchase and sale quantity, trading time, trading method, etc. of the underlying asset. Trading restrictions and maximum risk limits can be determined based on actual circumstances.

[0106] For example, based on the hedging type, the preset objective function and the transaction risk constraint, the optimal pricing mathematical model of the underlying asset can be expressed as follows:

[0107]

[0108] stδ∈H(δ)

[0109] Among them, ρ is the convex risk measurement function, Z is the liability corresponding to the underlying asset, δ k is the hedging strategy model corresponding to the underlying asset at time k, δ k+1 is the hedging strategy model at time k+1. The type of the hedging strategy model is the hedging type of the underlying asset. Different hedging types correspond to different hedging strategy models. The hedging strategy models of the same hedging type at different times may also be different. S k is the asset price of the underlying asset at time k, S k+1 is the price of the underlying asset at time k+1, C is the transaction cost function, and H is the transaction risk constraint. The transaction risk constraint may include a volatility change limit. The volatility change limit can be determined based on actual conditions and is not specifically limited in this embodiment. st is the abbreviation of subject to, which means "constrained to" in mathematical formulas. stδ∈H(δ) represents that δ is constrained to H(δ).

[0110] Step A33: Obtain a quantum neural network structure, and when the quantum neural network structure is in a preset approximation state, determine an initial quantum neural network model of the optimal pricing mathematical model based on the quantum neural network.

[0111] It should be noted that the term "quantum neural network structure" is used to describe a structure based on a quantum neural network model, which can include both classical and quantum layers. The initial quantum neural network model can be a representation of the optimal pricing mathematical model within the quantum neural network structure.

[0112] For example, the quantum neural network structure can be is a neural network with d0-dimensional input, d1-dimensional output, and m non-zero weights.

[0113] For any function exist When M→∞, F M →f. It can be a Lebesgue space, F M is a function sequence, M can also be expressed as a non-zero weight. When M approaches infinity, it belongs to F in M It can approach any function f. For tk The hedging strategy model at time δ k , t0 to t k The information set is {I0,…,I k}, I0 is the information at time t0, I k t k The information at the time, each piece of information in the information set may include the asset price, interest rate, and dynamic volatility of the underlying asset at time k, etc., which is not specifically limited in this embodiment.

[0114] Assume δ k Corresponding is predictable, then we know that there exists a function f such that δ k =f(I0,…,I k ), where δ k It can be a hedging strategy model at time k, It is t k The information flow at time t is the information flow of the underlying asset at t k The market data at that moment may be, for example, the transaction price of the underlying asset, news reports, etc. Function f may be a function corresponding to an option, and the corresponding I0,…,I in function f for different option types k different.

[0115] If we assume that δ k is Markovian (if δ k is Markov, then δ k Only depends on the information I at the current time k k , rather than information at all times), there exists δ k =f(I k ), and then there is Approximation f(I0,…,I k )=δ k . is the function sequence of parameter M at time k, F is the mapping of the neural network. When the parameter M approaches infinity, the input-output mapping F of the neural network can be approximated to the function f, so that the hedging strategy model δ can be derived using the information set I. k is the time, M is the parameter. In this embodiment, the quantum neural network structure in the preset approximation state may include: for any function exist When M→∞, F M →f, and existence Approximation f(I0,…,I k )=δ k .

[0116] Therefore, when the quantum neural network structure is in a preset approximation state, the mathematical model of optimal pricing can be represented in the quantum neural network structure as follows:

[0117]

[0118] in:

[0119]

[0120] Among them, Θ m,r(k+1),d for All parameters in the quantum neural network are all weights corresponding to the neurons in the quantum neural network; M (-Z) represents the optimal price. C is the transaction cost function.

[0121] Among them, Θ m,r(k+1),d Where m is a non-zero weight, d is the dimension of the quantum neural network output, r(k+1) is the dimension of the input, and k is the time. The ° in H°δ represents the Hadamard product. θ k is θ k The corresponding function sequence, θ is the neural network parameter, θ k is the neural network parameter at time k, F k is the function sequence at time k, δ k Represents the hedging strategy model at time k.

[0122] is a constraint condition, M is the complexity corresponding to the neural network; Z is a random number; T is the time when the option of the underlying asset expires.

[0123] In a feasible embodiment, step A30 further includes steps A31 to A33:

[0124] Step A31: Based on the initial quantum neural network model, determine the training hedging strategy model corresponding to each preset time step in the price path, and determine the training option pricing of the underlying asset;

[0125] It should be noted that the simulated price path can be input into the initial quantum neural network model, which can determine a training hedging strategy model corresponding to each preset time step in the price path. Each preset time step has a corresponding asset price, and each preset time step corresponds to a training hedging strategy model, which can be a training hedging strategy model corresponding to each asset price in the price path. A training hedging strategy model is a hedging strategy model corresponding to an asset price in the simulated price path.

[0126] After inputting the simulated price path into the initial quantum neural network model, the training option pricing of the underlying asset can also be output. The training option pricing is the option pricing corresponding to the simulated price path, which can be the option pricing corresponding to the last moment in the simulated price path.

[0127] Step A32: Calculate the training loss value corresponding to each training hedging strategy model and the training option pricing based on the optimal pricing mathematical model;

[0128] Step A33: updating the initial quantum neural network model based on the training loss value. When the training loss value is less than a preset loss threshold, the initial quantum neural network model is used as the preset quantum neural network model.

[0129] It should be noted that the simulated price path can be input into the initial quantum neural network model, which can determine a training hedging strategy model corresponding to each preset time step in the price path. Each preset time step has a corresponding asset price, and each preset time step corresponds to a training hedging strategy model, which can be a training hedging strategy model corresponding to each asset price in the price path. A training hedging strategy model is a hedging strategy model corresponding to an asset price in the simulated price path.

[0130] After inputting the simulated price path into the initial quantum neural network model, the training option pricing of the underlying asset can also be output. The training option pricing is the option pricing corresponding to the simulated price path, which can be the option pricing corresponding to the last moment in the simulated price path.

[0131] The training period pricing and each training hedging strategy model corresponding to the same price path can be substituted into the optimal pricing mathematical model to calculate the loss value corresponding to each preset time step. The smallest loss value is selected as the training loss value. The training loss value can be used to describe the preset accuracy of the initial quantum neural network model. The larger the training loss value, the lower the preset accuracy, and the smaller the training loss value, the higher the preset accuracy.

[0132] The preset loss threshold can be customized and is not specifically limited in this embodiment. When the training loss value is less than the preset loss threshold, it indicates that the initial quantum neural network model has a high prediction accuracy. The initial quantum neural network model can be used as the preset quantum neural network model, thereby facilitating the use of the preset quantum neural network model to predict the option price and hedging strategy model corresponding to the underlying asset.

[0133] Exemplarily, the simulated price path is input into the initial quantum neural network model to obtain a training hedging strategy model corresponding to each preset time step in the price path, as well as a training option pricing for the underlying asset. The training option pricing can be substituted into the optimal pricing mathematical model, and then the training hedging strategy model for each preset time step is substituted into the optimal pricing mathematical model to obtain a loss value corresponding to each preset time step. The smallest loss value is selected from the loss values ​​as the training loss value of the initial quantum neural network model. The initial quantum neural network model is updated using the training loss value to optimize the initial quantum neural network model. If the training loss value is less than a preset loss threshold, the initial quantum neural network model is used as the preset quantum neural network model. If the training loss value is greater than the preset loss threshold, the simulated price path is re-obtained and input into the initial quantum neural network model to obtain the corresponding training option pricing and each training hedging strategy model. The process then returns to step A32 until the training loss value is less than the preset loss threshold, thereby obtaining a trained preset quantum neural network model. Among them, updating the initial quantum neural network model based on the training loss value may be updating the model parameters in the initial quantum neural network model, for example, the structural parameters of the quantum circuit in the initial quantum neural network model, the weights of the quantum neurons, etc., which is not specifically limited in this embodiment.

[0134] This embodiment determines the training loss value through the optimal pricing mathematical model, so that the initial quantum neural network can be updated based on the minimum training loss value, thereby facilitating the improvement of the training efficiency of the initial quantum neural network and also facilitating the improvement of the accuracy of the initial quantum neural network.

[0135] For a better understanding of this embodiment, please refer to Figure 2 , Figure 2The following is a diagram of training the initial quantum neural network model. A simulated price path can be input into the initial quantum neural network model LZ. x0 to xT are the asset prices corresponding to the simulated price path. Each asset price is input into the classical layer of the initial quantum neural network model. The classical layer then inputs its output into the corresponding quantum layer. The quantum layer outputs the trained hedging strategy model δ0 to δT corresponding to each asset price. Based on each trained hedging strategy model, a training loss value is calculated. For example, each trained hedging strategy model can be input into the data model of the optimization problem to obtain a training loss value ρ(-z). The initial quantum neural network model can be updated based on ρ(-z). The classical layers in the initial quantum neural network include data normalization, a linear hidden layer, and a classical output layer. The corresponding functions of data normalization, linear hidden layer, and classical output layer can all be performed in the classical layers. The quantum layers include an input coding layer, a simulated circuit layer, and a measurement output layer. The corresponding functions of the input coding layer, simulated circuit layer, and measurement output layer can all be performed in the quantum layers. Among them, data standardization includes standardizing the simulated price path, the linear hidden layer can be used to extract features from the standardized data, the classic layer output layer can be used to convert the data output by the linear hidden layer into data that can be input into the input coding layer, and then transmitted to the input coding layer. The data output by the classic output layer is processed in the input coding layer and the simulated circuit layer, and the measured output layer outputs each trained hedging strategy model.

[0136] For a better understanding of this embodiment, refer to Figure 3 Taking the option type required for the underlying asset as Snowball option and the hedging strategy model type as Delta risk-neutral hedging as an example, the steps for determining the preset quantum neural network model are briefly illustrated: Among them, Snowball options include concepts such as knock-in and knock-out. Knock-in means that the option becomes effective only when the price of the underlying asset reaches or breaks through the preset barrier level; knock-out means that the option becomes effective only when the price of the underlying asset does not reach or break through the preset barrier level. The preset barrier level can be customized.

[0137] Step B10 sets market parameters and simulates the market: Market parameters corresponding to the underlying asset can be set first to simulate the market. For example, market parameters can include parameters such as the volatility and price drift rate of the underlying asset, which are not specifically limited in this embodiment. Step B20 sets the information set: An information set is set in the classical layer of the initial quantum neural network. The information set includes the price St of the underlying asset, the hedging strategy model of the previous cycle, a knock-in flag, and a knock-out flag. The knock-in flag can be used to determine whether an option is a knock-in, and the knock-out flag can be used to determine whether an option is a knock-out. For example, if the knock-in flag is empty, it can be considered not to have been knocked in, and if the knock-in flag is 1, it can be considered to have been knocked in; if the knock-out flag is empty, it can be considered not to have been knocked out, and if the knock-out flag is 1, it can be considered to have been knocked out. During the first cycle of the initial quantum neural network, the hedging strategy model of the previous cycle is not included in the information set.

[0138] Step B30: After setting the information set, determine whether the Snowball Option of the underlying asset has been knocked out. If the Snowball Option of the underlying asset has not been knocked out, proceed to step B40: classical feature extraction. The steps of classical feature extraction include: normalizing the information set and mapping the normalized information set to 2 using the linear hidden layer in the initial quantum neural network. n bit vector x=[x0,x1,…,x N-1 ], where N = 2 n , and normalize this vector again to get:

[0139]

[0140] Among them, x i is the vector corresponding to information i in the information set, |x| is the modulus of x, α i is the standardized x i The classic layer outputs a total of 2 n α i Then, step B50 is performed to perform quantum state amplitude coding. Quantum state amplitude coding can compress data to simplify data. Quantum state amplitude coding includes: converting each α output of the classical layer into i Converted to the input of the quantum layer:

[0141]

[0142] Among them, |ψ input > is the input quantum state, and |i> is the ground state of the quantum bit.

[0143] Using quantum entanglement to perform entanglement operations on adjacent qubits, the formula for performing entanglement operations on adjacent qubits can be expressed as:

[0144]

[0145] Among them, CNOT(i,i+1) means applying CNOT operation to the i-th quantum bit and the i+1-th quantum bit, CNOT gate (Control-NOT gate, double quantum logic gate), |ψ entangle >for|ψ input >The corresponding quantum entangled state;

[0146] Step B60: Quantum fully connected layer. The operation of the quantum fully connected layer includes: encoding parameters into the rotation gate and performing parameterized rotation gate operation on the entangled qubits:

[0147]

[0148] Among them, θ i is the rotation angle of the i-th revolving door in the quantum layer, R Y represents the rotation operation around the Y axis. The Y-axis rotation gate is a basic quantum gate used to change the state of the quantum bit. rot > is the |ψ that has been rotated entangle >.

[0149] Step B70: qubit measurement. The qubit measurement includes measuring the output of the qubit on the Pauli-Z basis in the initial quantum neural network model:

[0150]

[0151] In the field of quantum computing, the Pauli-Z basis is a quantum measurement method. Pauli matrices are fundamental operators in quantum mechanics, including X, Y, and Z. The Pauli-Z basis represents the measurement of quantum states along the Z axis. y can be expressed as the training option pricing for the output of the initial quantum neural network.

[0152] Step B80: Expiration: Determine whether the Snowball option of the underlying asset has expired. If the Snowball option has not expired, return to step B20. If the Snowball option has expired, execute step B90: Knock-in: Determine whether the Snowball option has been knocked in. If the Snowball option has not been knocked in or knocked out at expiration, execute step B100 to calculate the return based on the coupon rate. If the Snowball option has been knocked out, execute step B100 as well. If the Snowball option has been knocked out, the formula for calculating the return based on the coupon rate is:

[0153] payoff=S0*R*t

[0154] Where payoff is the payout, R is the coupon rate of the underlying asset, and t is the duration of the underlying asset, which is the time from the start time to the strike time. The start time refers to the starting point when Snowball options begin calculating payouts, and S0 is the price of the underlying asset at the start time. The coupon rate is the ratio of the dividend or coupon of the underlying asset to the current price of the underlying asset, and it affects the value of the option.

[0155] If the Snowball option expires without being knocked in or out, the formula for calculating the return based on the coupon rate is:

[0156] payoff=S0*R*T

[0157] Where T is the expiration time of the Snowball option.

[0158] If the Snowball option is knocked in but not knocked out at expiration, executing step B110 results in a loss or no gain. The formula for step B110 is:

[0159] payoff=min{S T -S0,0}

[0160] Among them, S T is the asset price of the snowball option of the underlying asset at expiration, when S T -When S0 is greater than 0, the training loss value is 0. T -When S0 is less than 0, the training loss value is S T -S0, and it also indicates that it is in a loss or no profit situation at this time.

[0161] After step B110 and step B100, step B120 is executed to optimize the parameters: the parameters of the initial quantum neural network are optimized based on the payoff; then step B130 is executed to determine whether a preset number of iterations has been reached: whether the number of times the initial quantum neural network has been optimized has reached a preset number of iterations. The preset number of iterations can be determined based on actual conditions. If the preset number of iterations has not been reached, the process returns to step B10. If the preset number of iterations is greater than or equal to the preset number of iterations, the training of the initial quantum neural network is terminated to obtain a preset quantum neural network.

[0162] In a feasible embodiment, after step S20, the method further includes:

[0163] Step S21: Based on the optimal option pricing data and the hedging strategy model, a hedging operation is performed on the underlying asset to obtain a hedging result;

[0164] Step S22: Optimize the preset quantum neural network model based on the hedging results.

[0165] It should be noted that a hedging operation is characterized as an operation performed based on a hedging strategy model output by a preset quantum neural network model. For example, the hedging strategy model may adjust the holdings of the underlying asset to a target value, and the hedging operation may be to adjust the holdings of the underlying asset to the target value. In other embodiments, the hedging operation may also include operations such as adjusting the underlying asset position to minimize risk and cost. Adjusting the underlying asset position means actively increasing or decreasing the amount of the underlying asset held.

[0166] The hedging result is characterized by the asset status of the underlying asset after the hedging operation is performed on the underlying asset, for example, it may include the profit or loss of the underlying asset, etc. The preset quantum neural network model can be optimized based on the hedging result to improve the accuracy of the preset quantum neural network model in predicting the optimal pricing data of the underlying asset's options and the hedging strategy model.

[0167] For a better understanding of this embodiment, please refer to Figure 4 and Figure 5 The overall process of this embodiment is briefly illustrated by way of example: the overall process of this embodiment may include price path simulation 1000, a mathematical model of the optimization problem 2000, a preset quantum neural network 3000, determination of optimal option pricing data for the underlying asset, and a hedging strategy model 4000.

[0168] In the price path simulation 1000, the simulation of the price path can be divided into modules for model selection, parameter selection, and numerical method selection. Model selection refers to the selection of a preset price simulation model for simulating the price path. The preset price simulation model can be selected from the BS model, the random volatility model, and the Jump Diffusion model (Jump Diffusion Model). The figure shows an example of the model, and this embodiment does not impose specific restrictions on this. The parameter selection module refers to determining the price simulation parameters required for the preset price simulation model and determining the preset time step of the preset price simulation model. The required price simulation parameters can be, for example, the price drift rate or the volatility rate. The numerical method module refers to rotating the preset numerical method used to generate the random path. For example, the preset numerical method can be selected from the Euler numerical method and the Monte Carlo method, or other numerical methods can be selected as the preset numerical method. The figure is only an example, and this embodiment does not impose specific restrictions on this.

[0169] In the mathematical model 2000 of the optimization problem, the construction of the mathematical model may include data simulation, determination of the corresponding objective function and determination of the corresponding constraints; the data simulation may be simulated by the BS model, the BS model may simulate the price path of the underlying asset, specifically, it may simulate the asset price of the underlying asset, simulate the risk-free interest rate and volatility, etc.; the objective function may be a convex risk measurement function, and the constraints may be transaction risk constraints, which may include volatility change limits, etc.

[0170] In the preset quantum neural network 3000, the preset quantum neural network may include classical feature extraction, quantum amplitude coding, quantum fully connected layer, quantum bit measurement and parameter optimization.

[0171] In determining the option pricing and hedging strategy model 4000 for the underlying asset, market data for the underlying asset is obtained, and the historical actual price path of the underlying asset can be determined from the market data. The preset quantum neural network model can output the option price and hedging strategy model of the underlying asset based on the historical actual price path. Hedging operations can be performed based on the hedging strategy model, and the preset quantum neural network model can be optimized based on the results of the hedging operations.

[0172] Reference Figure 6 , Figure 6 A schematic diagram of a module structure of an asset analysis device is shown in FIG. , and the device includes:

[0173] An acquisition module 10 is used to obtain the historical actual price path of the underlying asset in the asset market;

[0174] The analysis module 20 is configured to input the historical actual price path into a preset quantum neural network model, and output, through the preset quantum neural network model, option information of the underlying asset under the historical actual price path, wherein the option information includes optimal option pricing data and a hedging strategy model.

[0175] The asset analysis device provided in the present application adopts the asset option information generation method in the above-mentioned embodiment, aiming to solve the technical problem of low prediction accuracy of option pricing and hedging strategy models. Compared with the prior art, the beneficial effects of the asset option information generation method provided in the embodiment of the present application are the same as the beneficial effects of the asset option information generation method provided in the above-mentioned embodiment, and the other technical features in the asset analysis device are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here. The embodiment of the present application provides an electronic device, which can be a playback device, and the electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the asset option information generation method in the above-mentioned embodiment.

[0176] Reference below Figure 7 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0177] like Figure 7As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. Various programs and data required for the operation of the electronic device are also stored in the RAM 1004. The processing device 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although the figures show electronic devices with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented or have instead.

[0178] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication device 1009, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0179] The electronic device provided in this application utilizes the asset option information generation method described in the first embodiment above, aiming to address the technical issue of low prediction accuracy in option pricing and hedging strategy models. Compared to the prior art, the product flow data distribution provided in this embodiment offers the same beneficial effects as the asset option information generation method described in the first embodiment above. Other technical features of this asset analysis device are the same as those disclosed in the method described in the first embodiment above and are not further elaborated here. It should be understood that various aspects of this disclosure may be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in any one or more embodiments or examples. The foregoing description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed in this application are intended to be covered by the scope of protection of this application. Therefore, the scope of protection of this application shall be subject to the scope of protection of the claims set forth above. This embodiment provides a computer-readable storage medium having computer-readable program instructions stored thereon, which are used to execute the asset option information generation method described in the first embodiment above. The computer-readable storage medium provided in the embodiments of the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor devices, equipment or components, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable EPROM (Electrical Programmable Read Only Memory) or flash memory, an optical fiber, a portable compact disk CD-ROM (compact disc read-only memory), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution device, device or component. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof. The above-mentioned computer-readable storage medium may be contained in an electronic device; or it may exist separately without being assembled into an electronic device.The computer-readable storage medium carries one or more programs. When the one or more programs are executed by an electronic device, the electronic device is enabled to: obtain the historical actual price path of the underlying asset in the asset market; input the historical actual price path into a preset quantum neural network model, and output option information of the underlying asset under the historical actual price path through the preset quantum neural network model, wherein the option information includes option optimal pricing data and a hedging strategy model.

[0180] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a LAN (local area network) or a WAN (wide area network), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0181] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the devices, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based device that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0182] The modules described in the embodiments of the present disclosure may be implemented in software or in hardware. The name of the module does not, in some cases, limit the unit itself. The computer-readable storage medium provided in this application stores computer-readable program instructions for executing the above-mentioned method for generating option information for the asset, and is intended to solve the technical problem of low prediction accuracy of option pricing and hedging strategy models. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the embodiments of this application are the same as the beneficial effects of the method for generating option information for the asset provided in the above-mentioned embodiments, and will not be elaborated here.

[0183] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for generating option information of an asset as described above. The computer program product provided by the present application is intended to solve the technical problem of low prediction accuracy of option pricing and hedging strategy models. Compared with the prior art, the beneficial effects of the computer program product provided by the embodiment of the present application are the same as the beneficial effects of the method for generating option information of an asset provided by the above embodiment, and will not be described in detail here. The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the description and drawings of this application, or directly or indirectly applied in other related technical fields, are also included in the patent processing scope of the present application.

Claims

1. A method for generating asset option information, characterized in that: The method includes: Obtain the historical actual price path of the underlying asset in the asset market; The historical actual price path is input into a preset quantum neural network model, and the preset quantum neural network model outputs option information of the underlying asset under the historical actual price path, wherein the option information includes option optimal pricing data and a hedging strategy model.

2. The method according to claim 1, wherein The method includes: Obtaining historical market data for the underlying asset; simulating a price path of the underlying asset based on the historical market data; An initial quantum neural network model is obtained, and the initial quantum neural network model is trained according to the price path to obtain a trained preset quantum neural network model.

3. The method according to claim 2, wherein The step of simulating the price path of the underlying asset based on the historical market data includes: Get the preset price simulation model and the preset time step; Obtaining price simulation parameters required by the preset price simulation model from the historical market data; Configuring the price simulation parameters and the preset time step in the preset price simulation model; Based on a preset price simulation model and a preset random numerical method, the price path of the underlying asset is simulated.

4. The method according to claim 3, wherein The step of simulating the price path of the underlying asset based on a preset price simulation model and a preset random value method includes: Generate a random path for the underlying asset based on a preset random value method; The preset price simulation model is used to generate the asset price of the underlying asset at each preset time step in the random path, thereby obtaining the price path of the underlying asset.

5. The method according to claim 2, wherein The step of obtaining the initial quantum neural network model includes: Get the hedge type of the underlying asset; Constructing an optimal pricing mathematical model for the underlying asset based on the hedge type, a preset objective function, and the transaction risk constraints of the underlying asset; A quantum neural network structure is obtained, and when the quantum neural network structure is in a preset approximation state, an initial quantum neural network model of the optimal pricing mathematical model is determined based on the quantum neural network.

6. The method according to claim 5, wherein The step of training the initial quantum neural network model according to the price path to obtain a trained preset quantum neural network model includes: Determining, based on the initial quantum neural network model, a training hedging strategy model corresponding to each preset time step in the price path, and determining a training option pricing for the underlying asset; Calculating, based on the optimal pricing mathematical model, a training loss value corresponding to each of the training hedging strategy models and the training option pricing; The initial quantum neural network model is updated based on the training loss value, and when the training loss value is less than a preset loss threshold, the initial quantum neural network model is used as a preset quantum neural network model.

7. The method according to claim 1, wherein After the step of inputting the historical actual price path into a preset quantum neural network model and outputting the option information of the underlying asset under the historical actual price path through the preset quantum neural network model, the method further includes: Based on the optimal option pricing data and the hedging strategy model, performing a hedging operation on the underlying asset to obtain a hedging result; Based on the hedging result, the preset quantum neural network model is optimized.

8. An asset option information generation device, characterized in that: The device comprises: The acquisition module is used to obtain the historical actual price path of the underlying asset in the asset market; An analysis module is configured to input the historical actual price path into a preset quantum neural network model, and output, through the preset quantum neural network model, option information of the underlying asset under the historical actual price path, wherein the option information includes optimal option pricing data and a hedging strategy model.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method for generating asset option information according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, on which is stored a program for implementing a method for generating asset option information. The program for implementing a method for generating asset option information is executed by a processor to implement the steps of the method for generating asset option information as described in any one of claims 1 to 7.

Images