CN116299853A - Curved waveguide and curved waveguide design method based on genetic algorithm - Google Patents

Abstract

The invention provides a curved waveguide and a curved waveguide design method based on a genetic algorithm, which randomly generates an initial population, wherein the initial population comprises a plurality of individuals, genes of each individual are encoded according to parameters of two curves forming the curved waveguide, parameters of a plurality of curves dividing an etching area and etching parameters of the etching area, the geometric shape and layout of each individual are determined according to the genes of each individual in the population, a device model of each individual is obtained, and the transmissivity and crosstalk of each individual are obtained according to the device model of each individual; calculating the fitness of each individual according to the transmissivity and crosstalk of each individual, and finding out the individual with the highest fitness in the population as the optimal individual; judging whether the iteration termination condition is met, if so, outputting an optimal individual, ending the iteration, if not, updating the iteration times, generating a next generation population, and repeating the operation until the optimal individual is output. The invention improves the design efficiency of the curved waveguide.

Description

Curved waveguide and design method of curved waveguide based on genetic algorithm

technical field

The invention relates to the technical field of curved waveguides, in particular to a curved waveguide and a genetic algorithm-based design method for the curved waveguide.

Background technique

Bending waveguides are mainly used in on-chip photonic/optoelectronic integration, and more recently in the fields of optical communication and optical computing. In order to realize on-chip photonic integration, the wiring of the waveguide must be considered. In order to complete any wiring, the waveguide must be bent. However, when light passes through a curved waveguide, the curvature of the waveguide caused by the bending will cause mode mismatch and distortion, and spurious modes different from the input mode will appear, resulting in loss and crosstalk. Especially in the case of on-chip photonic integration, the superposition of loss and crosstalk of multiple curved waveguides will exceed the tolerance of practical applications.

In order to reduce the loss and crosstalk of curved waveguides, a curved waveguide based on artificial micro-nanostructures that can be applied to on-chip photonic integration is proposed. For curved waveguides based on artificial micro-nano structures, the most important thing, and what their performance depends on, is the geometric structure of their micro-nano structures (also called design layout). The geometric structure is generally determined by several parameters. Therefore, how to obtain the values of these parameters has become the main technical problem of curved waveguide design.

Contents of the invention

In view of this, the present invention provides a curved waveguide and a method for designing the curved waveguide based on a genetic algorithm.

In the first aspect, the present invention provides a curved waveguide, wherein a micro-nano structure is formed on the curved waveguide, and the micro-nano structure is any one of the following: a plurality of nano-slots distributed along the direction of light propagation; Multiple nanoslots distributed in the direction of propagation or nanoholes of arbitrary shape in the curved waveguide region.

Optionally, the shape of the two curves constituting the curved waveguide is a hyperelliptic curve.

Optionally, when the micro/nano structure is a plurality of nanogrooves distributed along the direction of light propagation, the shape of the two curves constituting each of the nanogrooves is a hyperelliptic curve.

In a second aspect, the present invention provides a method for designing a curved waveguide based on a genetic algorithm, including:

Step 1), randomly generate an initial population, including a plurality of individuals, each individual's gene is based on the parameters of the two curves that constitute the curved waveguide, the parameters of the multiple curves that divide the etched area, and the parameters of the etched area. Etching parameters are encoded, wherein the etched area is used to form micro-nano structures on the curved waveguide;

Step 2), according to the gene of each individual in the population, determine the geometric shape and layout of each individual, obtain the device model of each individual, and obtain each device model according to the device model of each individual The transmittance and crosstalk of each of the individuals;

Step 3), according to the transmittance and crosstalk of each individual, calculate the fitness of each individual respectively, and find out the individual with the highest fitness in the population as the optimal individual;

Step 4), judging whether the iteration termination condition is reached, if so, then output the optimal individual, and end the iteration, if not, update the number of iterations, and perform step 5);

Step 5), generate the next generation population through selection operation, crossover operation and mutation operation, return to step 2).

Optionally, the shape of the two curves constituting the curved waveguide is a hyperelliptic curve;

The analytical formula of hyperelliptic curve is:

Among them, n, a, and b are all positive numbers, and ∣∣ represents the absolute value.

Optionally, when the micro-nano structure is a plurality of nano-grooves distributed along the direction of light propagation, the shape of the curve dividing the etching area is a hyperelliptic curve;

The genes of each individual are coded according to the two hyperelliptic curves constituting the curved waveguide, the parameter n of multiple hyperelliptic curves dividing the etched area, and the etching parameters of the etched area.

Optionally, the etching parameters of the etching area are a series of binary numbers, each binary number is used to indicate whether the corresponding etching area is etched, 0 indicates no etching, and 1 indicates etching.

Optionally, the selection operation includes: according to the elite retention strategy, retaining the best individuals in the population directly enters the next generation population; using the tournament method for the current population to select individuals with higher fitness as the parent of the next generation population and the parent;

The crossover operation includes: according to the set crossover point and crossover rules, crossover the chromosomes of the selected father and mother, and the obtained individuals enter the next generation population;

The mutation operation includes: mutating the chromosomes of the next generation population according to the set number of mutated individuals and the number of mutated digits.

In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and operable on the processor. When the processor executes the computer program, the above-mentioned genetic algorithm-based Curved waveguide design method.

In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned curved waveguide design method based on a genetic algorithm is implemented.

A curved waveguide provided by an embodiment of the present invention, by etching micro-nano structures on the curved waveguide, introducing refractive index perturbation to regulate the mode, compensate the mode mismatch, and suppress the mode distortion to achieve low loss and low Crosstalk, small size (small bend radius) and support for multiple modes of transmission. In addition, the curved waveguide design method based on the genetic algorithm is a reverse design method, which can quickly determine the etching parameters of the micro-nano structure according to the design requirements, and obtain the design layout of the curved waveguide. Moreover, it has good versatility and can be applied to the design of other similar micro-nano optical devices.

Description of drawings

FIG. 1 is a schematic structural diagram of a curved waveguide according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of dividing etching regions according to an embodiment of the present invention;

3 is a schematic flow chart of a method for designing a curved waveguide based on a genetic algorithm according to an embodiment of the present invention;

FIG. 4 is a partial design result of applying the curved waveguide design method according to the embodiment of the present invention;

FIG. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is only some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

It should be noted that the terms "first" and "second" in the description and claims of the present application and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It should be understood that the data so used may be interchanged under appropriate circumstances for the embodiments of the application described herein. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or device comprising a sequence of steps or elements is not necessarily limited to the expressly listed instead, may include other steps or elements not explicitly listed or inherent to the process, method, product or apparatus.

Some embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. In the case of no conflict, the following embodiments and features in the embodiments can be combined with each other.

An embodiment of the present invention provides a curved waveguide. A micro-nano structure is formed on the curved waveguide. The micro-nano structure is any one of the following: a plurality of nano-grooves distributed along the direction of light propagation, and a plurality of nano-grooves distributed perpendicular to the direction of light propagation. Multiple nanoslots or nanoholes of arbitrary shape in the curved waveguide region.

In one embodiment, the material of the curved waveguide is Si-SiO ₂ -Si SOI (silicon-on-insulator) system, and the top silicon of the curved waveguide is formed with the micro-nano structure. Nanogrooves are obtained by shallow etching on the top silicon, and nanopores are obtained by full etching on the top silicon.

The curved waveguide is composed of two specific curves, and the curves are arbitrary curves that can be determined by certain parameters and analytical formulas, such as circular arcs, Bezier curves, Euler curves, superelliptic curves, etc. Fig. 1 is a schematic structural diagram of a curved waveguide according to an embodiment, as shown in Fig. 1 , the upper diagram in Fig. 1 is a top view of the curved waveguide, and the lower diagram is an elevation view of the curved waveguide. The shape of the two curves that make up the curved waveguide is a hyperelliptic curve.

The analytical formula of hyperelliptic curve is:

Among them, n, a, and b are all positive numbers, and ∣∣ represents the absolute value. The values of a and b are related to the bending radius and have no special meaning, they are a constant. A meaningful value is the coefficient n, and the value of n determines the shape of the curve. For example, when n=2, the curve is a circular arc. When n=3, the curve is similar to the Euler curve commonly used in multimode curved waveguide design. Simply put, hyperelliptic curves can contain some special curves except free curves.

In addition, when the micro-nano structure is a plurality of nano-grooves distributed along the direction of light propagation, preferably, the shape of the two curves constituting each nano-groove is a hyperelliptic curve. In Fig. 1, the parameter n of each hyperelliptic curve can be the same or different.

The thickness of the top silicon depends on the required mode and the process level. Only TE mode is supported. The thickness of the top silicon is generally selected as 220nm. Both TE mode and TM mode are supported. The thickness of the top silicon is 340nm, and the working wavelength is the communication wavelength of 1550nm. The etching depth of the nano-groove is generally determined by the process level and requirements. Generally, only TE is supported. If the top silicon is 220nm thick, the nano-groove will be shallower. If it supports TE and TM, the nano-groove will be deeper if the top silicon is 340nm thick.

The embodiment of the present invention optimizes the structure of the curved waveguide to a certain extent. By etching micro-nano structures on the curved waveguide, for example, shallow etching of nano-grooves along the direction of light propagation, and introducing perturbation of the refractive index to regulate the mode, the Mode mismatch is compensated and mode distortion is suppressed to achieve low loss, low crosstalk, small size (small bending radius) and support for multiple modes of transmission.

Specifically, the optimization methods include: curve shape optimization and shallow etching of nano-grooves.

First, the meaning of curve shape optimization refers to selecting a special curve as the curved waveguide curve. The curved waveguide curve used in the embodiment of the present invention is a hyperelliptic curve. Referring to the analytical formula of the hyperelliptic curve given above, in order to ensure symmetry, a=b can be taken here.

In addition, in general, the width of the curved waveguide is equal to that of the straight waveguide. For the hyperelliptic curve used in this embodiment, if the coefficients n of the two curves constituting the curved waveguide are equal, they should also be of equal width. However, in order to better support high-order modes and reduce losses, it is also possible to choose to set the coefficient n of the two curves constituting the curved waveguide to be variable. Whether the coefficient n is the same or different is determined by the algorithm, which will be introduced in detail later.

Second, shallow etching of nano-grooves means etching and forming multiple nano-grooves on the top silicon of the curved waveguide. When a plurality of nanogrooves are distributed along the direction of light propagation, the shape of the two curves constituting each nanogroove also adopts hyperelliptic curves.

Further, in this embodiment, the formation method of the nano-groove is: first divide a plurality of etching regions on the curved waveguide, hereinafter referred to as the nano-groove etching region, and then determine whether the nano-groove etching region at each position is is etched. The following is a detailed introduction.

First, the division method of the etched area of the nano-groove depends on the feature size.

Examples are as follows:

Assuming that the curved waveguide is required to support 4 TE modes, the thickness of the top silicon is selected to be 220nm, the waveguide width is determined to be 1500nm by the dispersion curve, and the characteristic size is 100nm. Referring to Figure 2, when dividing the nano-groove etching area, the method of equal division is adopted, and the connection between the straight waveguide and the curved waveguide is divided into 15 sections (circled in Figure 2), and there is one section every 100nm. This 100nm is the characteristic size. It is also the minimum width of the next nanoslot. Along the direction of light propagation, by connecting the endpoints divided by the incident end and the outgoing end (there are 14 endpoints at both ends), the corresponding 15 nano-groove etching regions can be divided.

There are a total of 16 curves, including two curves forming the curved waveguide and 14 curves dividing the etched area of the nanoslot. These 16 curves are hyperelliptic curves, and the coefficient n of each hyperelliptic curve is not necessarily the same, but can be changed. Suppose the coefficient of the first curve is n, the gradient of the curve is dn, the coefficient of the second curve is n+dn, similarly, the coefficient of the kth curve is n+k*dn. n and dn are determined by the algorithm. When dn is 0, it is a waveguide with equal width, and when dn is not 0, it is a waveguide with gradually changing width. The narrowest width location is where it connects to the straight waveguide.

After the division of the etched area of the nano-groove is completed, an algorithm determines whether the etched area of the nano-groove at each position is to be etched.

To put it simply, 14 hyperelliptic curves are used on the curved waveguide to divide 15 nanogroove etching regions along the direction of light propagation. The elliptic curve determines that part of the etched area of the nano-groove is etched, and part of the etched area of the nano-groove is retained, and finally a plurality of nano-grooves are formed.

On the other hand, an embodiment of the present invention provides a method for designing a curved waveguide based on a genetic algorithm. As shown in FIG. 3 , the method includes steps S301-S305.

Step S301, randomly generate an initial population, including multiple individuals, each individual's gene is encoded according to the parameters of the two curves that constitute the curved waveguide, the parameters of the multiple curves that divide the etching area, and the etching parameters of the etching area , where the etched area is used to form micro-nano structures on the curved waveguide;

Step S302, according to the gene of each individual in the population, determine the geometry and layout of each individual, obtain the device model of each individual, and obtain the transmittance and crosstalk of each individual by solving according to the device model of each individual;

Step S303, calculate the fitness of each individual according to the transmittance and crosstalk of each individual, and find the individual with the highest fitness in the population as the optimal individual;

Step S304, judging whether the iteration termination condition is met, if so, output the optimal individual, and end the iteration, if not, update the number of iterations, and execute step S305;

Step S305, generate the next generation population through selection operation, crossover operation and mutation operation, return to step S302.

Each step is introduced below.

In step S301, an initial population is randomly generated, and the population has a plurality of individuals, assuming that the population has 40 individuals, and one individual represents a designed curved waveguide, that is, there are 40 different curved waveguides. Each individual's genes are encoded according to the parameters of the two curves that make up the curved waveguide, the parameters of the multiple curves that divide the etched area, and the etching parameters of the etched area.

It is specifically expressed as 40 different binary strings (that is, individual genes or chromosomes), such as individual 1: 010010..., individual 2: 101010.... Each binary string represents a parameter of the curved waveguide.

In one embodiment, the two curves constituting the curved waveguide are hyperelliptic curves, and the analytical formula of the hyperelliptic curve is:

Among them, n, a, and b are all positive numbers, and ∣∣ represents the absolute value.

Further, the micro-nano structure is any one of the following: a plurality of nanogrooves distributed along the direction of light propagation, a plurality of nanogrooves distributed perpendicular to the direction of light propagation, or nanoholes of arbitrary shape. When the micro-nano structure is a plurality of nano-grooves distributed along the direction of light propagation, the multiple curves dividing the etching area also adopt hyperelliptic curves;

Correspondingly, the genes of each individual are encoded according to the two hyperelliptic curves constituting the curved waveguide, the parameter n of multiple hyperelliptic curves dividing the etched area, and the etching parameters of the etched area.

Here is an example. Assuming that it is divided into 15 etching areas, two hyperelliptic curves that constitute the curved waveguide and 14 hyperelliptic curves that divide the etching area, there are 16 curves in total, and the genetic code includes the parameter n of the 16 hyperelliptic curves and the etching Etching parameters for the region. If there is an association between these 16 hyperelliptic curves, for example, the gradient is dn, then the genetic code includes n and dn. Suppose the coefficient of the first curve is n, the gradient of the curve is dn, the coefficient of the second curve is n+dn, similarly, the coefficient of the kth curve is n+k*dn. n and dn are determined by the algorithm. When dn is 0, it is a waveguide with equal width, and when dn is not 0, it is a waveguide with gradually changing width. The narrowest width location is where it connects to the straight waveguide.

As for the etching parameters of the etching area, a series of binary numbers are used, and each binary number is used to indicate whether the corresponding etching area is etched, 0 indicates no etching, and 1 indicates etching.

The number of bits for an individual gene is related to the number of curved waveguide parameters and the number of character bits for each parameter. For example: the curved waveguide has 40 parameters, and each parameter is composed of 4 characters, such as 1010, then the total number of bits of an individual gene is 160 bits. The specific gene coding rules can be defined by yourself.

In addition, the algorithm itself has added restrictions. If there are identical individuals, only one of them will be kept, and new and different individuals will be regenerated, that is, the genes of each individual are different.

In step S302, the gene is decoded to realize the conversion between the binary string of the individual gene and the parameters, determine the geometry and layout of each individual, obtain the device model of each individual, and according to the device model of each individual Use the finite element (FEM) solution software COMSOL to solve the electromagnetic field of the curved waveguide, and obtain the transmittance and crosstalk of each individual (with the help of the interface between Matlab and COMSOL, the code is automatically generated and the model is automatically solved). Transmittance and crosstalk represent the performance of the curved waveguide. The higher the transmittance, the better, and the lower the crosstalk, the better.

In step S303, combining the transmittance and crosstalk of each individual, an appropriate fitness function is selected to characterize the performance of the curved waveguide. This application does not specifically limit the fitness function.

In step S304, it is judged whether the iteration termination condition is reached, and the termination condition can set two conditions: whether to converge, or whether to reach the target. If any condition is met, the optimized individual is output as the result. If the two conditions are not met, proceed to the next step to generate the next generation population.

Specifically, examples are as follows:

Judging whether it is converged: Assume that the optimal individual fitness of the 50th generation is 0.901, and the optimal individual fitness of the 40th generation ten generations ago is 0.9. Ten generations are computed, but the improvement is less than 1%, it is considered converged. If we continue to optimize, the performance improvement will be small, and the cost performance will be extremely low. At this point, the calculation is stopped.

Judging whether the goal is reached: For example, if the individual performance requirement is a fitness of 0.9 or more, it will end when it is reached.

In step S305, the next generation population is generated through selection operation, crossover operation and mutation operation. Since the individuals in the current population do not meet the requirements, it is necessary to generate the next generation population by selecting excellent individuals and crossing their genes to generate individuals in the next generation population. At the same time, introduce variation to ensure the genetic diversity of the population and jump out of the optimization trap.

Selection operation: In the embodiment of the present invention, according to the elite retention strategy, the best individual in the reserved population directly enters the next generation population; the current population adopts the tournament method to select the individual with the highest fitness as the male parent and female parent of the next generation population, to Guarantee the retention of excellent genes.

Elite retention strategy: Simply put, the best individuals in the current population are directly retained into the next generation population. For example, the optimal individual gene of the current population is 10001000, and there are 40 individuals in the next generation population, then there must be one individual whose gene is 10001000 in the next generation, and the genes of the other 39 individuals are produced by selection, crossover, and mutation. To put it simply, the best individuals are recommended to the next generation population to prevent excellent individuals from disappearing due to crossover or mutation, and also to ensure that the more optimized the best individuals, the worse they will be.

After retaining the optimal individual, there are still 39 individual genes to be generated. Each gene is formed by the crossover of the genes of the parents. Then we have to choose 39 pairs of parents, and different selection strategies will have a certain impact on the results. The embodiment of the present invention adopts the tournament method to select individuals with higher fitness as the male parent and female parent of the next generation population.

For example, out of 40 individuals, 10 are randomly selected, and 1-10 are selected to compare their fitness. Among the 10, the best is No. 1, and No. 1 is the male parent. The mother parent was obtained in the same way. It is obvious that the higher the fitness, the better the individual, and the greater the probability of being selected as a parent.

It should be noted that in actual selection, there may be situations where the individual is so good that both parents are the individual in a certain selection. This kind of mutation operation will be involved later, which will be explained later.

Crossover operation: the embodiment of the present invention crosses the chromosomes of the selected father and mother according to the set crossover point and crossover rules, and the obtained individuals enter the next generation population. That is, both individuals are excellent, but it is not known which parameter makes it perform better. So just mix the genes of two individuals and see if it's better or worse.

An example is as follows: the paternal chromosome is 100100100100, the maternal chromosome is 110110110110, and the chromosome is 12 bits. A number is randomly selected, and if it is 6, the intersection point is the sixth point. The first six chromosomes of the offspring are the first six chromosomes of the father. Six digits, the last six chromosomes of the offspring are the last six digits of the mother (the reverse is also possible, set freely). The offspring chromosome is 100100110110.

It should be noted that the intersection point is not necessarily 1, and is generally related to the length of the chromosome. Longer chromosomes generally have more intersection points. Because the chromosome is too long and there are too many parameters, only one bit of crossover will change too little, and the evolution rate is too slow. For example, in one embodiment, when there are 160 chromosomes, 3 intersection points may be selected.

Mutation operation: The mutation operation is to change a certain chromosome of an individual randomly selected to generate a new chromosome after the crossover is completed and a new set of chromosomes is generated. The embodiments of the present invention mutate the chromosomes of the next generation population according to the set number of mutated individuals and the number of mutated digits.

For example, there are 40 individuals in the next-generation population, and the chromosome of an optimal individual in the previous generation is retained, and 39 chromosomes are generated by selection operation and crossover operation. For example, two chromosomes 1 and 20 are randomly selected. Chromosomes 1 and 20 are 11111111 and 00000000 respectively. Then take a bit for each of them, for example, the No. 1 chromosome takes the third place, mutates (0 turns into 1, 1 turns into 0), and the new individual is 11011111. The No. 20 individual takes the fourth place, and after the mutation, the new individual is 11011111. The individual is 00010000.

The most important role of mutation is to increase the diversity of the population and jump out of the local optimal solution.

For example, a certain generation suddenly has a very good super individual (the fitness of other individuals is generally around 0.2, and the fitness of this super individual is 0.7), because the super individual will always be selected in the selection operation, and it may converge quickly around the super entity. However, this individual may be a local optimal solution, that is, there is still an individual with a fitness of 0.9, but now it is difficult to jump out by crossover (too similar chromosomes) when it is trapped near this individual. So it needs to be mutated.

It should be noted that the proportion of mutated individuals in the entire population should not be too high, and when the proportion is too high, the genetic algorithm will approach random search. At the same time, the ratio of the number of mutations to the number of chromosomes should not be too high. The two ratios depend on the number of depopulated individuals and the number of chromosomes respectively.

Restrictions are also added to the algorithm. Only one individual with the same chromosome will be retained, and the vacancy after deleting the duplicate individual will be filled by random mutation of the optimal individual of the previous generation (the optimal individual left by the elite retention strategy). In this way, it can ensure that there are no repeated individuals in each generation, and effectively search around the optimal individual.

After the next-generation population is generated, return to step S302, and perform this repeatedly until the iteration termination condition is reached, and the optimal individual is output.

The optimal individual is obtained, and the micro-nano structure is etched on the curved waveguide according to the parameters of the optimal individual, and a high-performance curved waveguide can be obtained. Fig. 4 is a design result of a series of high-performance multimode curved waveguides obtained by applying the design method of the embodiment of the present invention.

A method for designing a curved waveguide based on a genetic algorithm provided by an embodiment of the present invention is a reverse design method, which can quickly determine the etching parameters of the curved waveguide according to design requirements, and obtain the design layout of the curved waveguide. Moreover, it has good versatility and can be applied to the design of other similar micro-nano optical devices. This application is intended for manufacturers in the fields of on-chip photonics/optoelectronic integrated chips, silicon photonics, optical computing, and optical communications. This application has certain application prospects.

FIG. 5 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. As shown in FIG. 5 , the electronic device may include: a processor (processor) 501 , a memory (memory) 502 and a communication bus 503 , where the processor 501 and the memory 502 communicate with each other through the communication bus 503 . The processor 501 can execute the program instructions in the memory 502, so as to implement the method for designing a curved waveguide based on a genetic algorithm provided in the foregoing embodiments.

On the other hand, an embodiment of the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the curved waveguide design based on the genetic algorithm provided by the above-mentioned embodiments is realized. method.

Through the above description of the implementations, those skilled in the art can clearly understand that each implementation can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic discs, optical discs, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (10)

1. A curved waveguide, characterized in that a micro-nano structure is formed on the curved waveguide, and the micro-nano structure is any one of the following: a plurality of nano-grooves distributed along the direction of light propagation, perpendicular to the light propagation Multiple nanoslots distributed in different directions or nanoholes of arbitrary shape in the curved waveguide region. 2. The curved waveguide according to claim 1, wherein the shape of the two curves constituting the curved waveguide is a hyperelliptic curve. 3. The curved waveguide according to claim 2, characterized in that, when the micro-nano structure is a plurality of nano-grooves distributed along the direction of light propagation, the shape of two curves constituting each of the nano-grooves is a hyperelliptic curve. 4. A curved waveguide design method based on genetic algorithm, characterized in that said method comprises: Step 1), randomly generate an initial population, including a plurality of individuals, each individual's gene is based on the parameters of the two curves that constitute the curved waveguide, the parameters of the multiple curves that divide the etched area, and the parameters of the etched area. Etching parameters are encoded, wherein the etched area is used to form micro-nano structures on the curved waveguide; Step 2), according to the gene of each individual in the population, determine the geometric shape and layout of each individual, obtain the device model of each individual, and obtain each device model according to the device model of each individual The transmittance and crosstalk of each of the individuals; Step 3), according to the transmittance and crosstalk of each individual, calculate the fitness of each individual respectively, and find out the individual with the highest fitness in the population as the optimal individual; Step 4), judging whether the iteration termination condition is reached, if so, then output the optimal individual, and end the iteration, if not, update the number of iterations, and perform step 5); Step 5), generate the next generation population through selection operation, crossover operation and mutation operation, return to step 2). 5. The curved waveguide design method based on genetic algorithm according to claim 4, wherein the shape of the two curves forming the curved waveguide is a hyperelliptic curve; The analytical formula of hyperelliptic curve is:

Among them, n, a, and b are all positive numbers, and ∣∣ represents the absolute value. 6. The curved waveguide design method based on genetic algorithm according to claim 5, wherein when the micro-nanostructure is a plurality of nano-grooves distributed along the direction of light propagation, the curve dividing the etched area The shape is a hyperelliptic curve; The genes of each individual are coded according to the two hyperelliptic curves constituting the curved waveguide, the parameter n of multiple hyperelliptic curves dividing the etched area, and the etching parameters of the etched area. 7. The method for designing a curved waveguide based on a genetic algorithm according to any one of claims 4 to 6, wherein the etching parameters of the etching region are a string of binary numbers, and each binary number is used to represent a corresponding Whether the etched area is etched, 0 means not etched, 1 means etched. 8. The curved waveguide design method based on genetic algorithm according to claim 4, characterized in that, the selection operation comprises: according to the elite retention strategy, the best individual in the reserved population directly enters the next generation population; for the current population Using the championship method to select individuals with higher fitness as the male and female parents of the next generation population; The crossover operation includes: according to the set crossover point and crossover rules, crossover the chromosomes of the selected father and mother, and the obtained individuals enter the next generation population; The mutation operation includes: mutating the chromosomes of the next generation population according to the set number of mutated individuals and the number of mutated digits. 9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and operable on the processor, characterized in that, when the processor executes the computer program, any of claims 4 to 8 can be realized. A curved waveguide design method based on genetic algorithm. 10. A non-transitory computer-readable storage medium, on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the curved waveguide based on a genetic algorithm according to any one of claims 4 to 8 is realized design method.

Images