Illumination Scheme Design Method and Device Based on Artificial Fish School and Differential Evolution

Abstract

The present application provides a lighting scheme design method and device based on artificial fish swarm and differential evolution. The method includes determining the reference values of lighting design indicators and the pre-information of the target lighting space; initializing the population to be optimized and algorithm parameters; each individual in the population to be optimized includes a set of decision variables for determining the lighting scheme; iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm in sequence until the number of optimization generations reaches the termination threshold, and finally selecting a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space. In the inverse design process of the lighting scheme, the present invention combines the differential evolution and the artificial fish swarm algorithm. Compared with the method using a single algorithm, it has a faster convergence speed and is not easily trapped in local optimal solutions. Therefore, a lighting design scheme meeting the requirements can be obtained faster based on the present invention.

Description

Technical Field

[0001] The present invention relates to the technical field of lighting scheme design, and particularly to a lighting scheme design method and device based on artificial fish swarm and differential evolution. Background Art

[0002] Currently, the common method for designing a lighting scheme is to rely on the experience estimation of lighting designers to obtain an initial scheme, and then obtain the final scheme through the process of repeatedly "simulating - adjusting - re - simulating - re - adjusting" the initial scheme. This process is time - consuming and laborious, and it is not easy to meet multiple index requirements simultaneously.

[0003] Another lighting scheme design method is inverse design, that is, using the expected performance of the system as input and the design parameters as output, and using heuristic algorithms to generate a feasible design scheme. Common heuristic algorithms for solving lighting inverse design problems include genetic algorithm, differential evolution algorithm, particle swarm algorithm, ant colony algorithm, annealing algorithm, artificial fish swarm algorithm, etc. However, the above algorithms have disadvantages such as slow convergence speed and easy to fall into local optimal solutions to varying degrees. Therefore, inverse design based on the above algorithms still has problems such as low efficiency and the final output lighting scheme cannot accurately meet the requirements. Summary of the Invention

[0004] Aiming at the above - mentioned shortcomings of the prior art, the present invention provides a lighting scheme design method and device based on artificial fish swarm and differential evolution to provide an efficient and accurate lighting scheme design method.

[0005] The first aspect of the present application provides a lighting scheme design method based on artificial fish swarm and differential evolution, including:

[0006] Determine the reference value of lighting design indicators and the pre - information of the target lighting space; wherein, the pre - information includes the size information, shape information, lamp layout range information and lighting work surface of the target lighting space;

[0007] Initialize the population to be optimized and algorithm parameters; wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme;

[0008] Iteratively optimize the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset mode threshold;

[0009] Perform mode switching, and iteratively optimize the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the number of optimization generations reaches a preset termination threshold; wherein, the mode switching is only performed once when the number of optimization generations reaches the mode threshold; the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population to be optimized; the unimproved individuals refer to the individuals that have not been updated during the iterative optimization based on the differential evolution algorithm.

[0010] Select a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space; wherein, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of the individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables and variable values included in the individual and the pre-information of the target lighting space.

[0011] Optionally, the iterative optimization of the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset mode threshold includes:

[0012] For each individual in the population to be optimized, perform mutation, crossover, and selection operations in sequence.

[0013] Among them, the mutation operation is represented by the following formula:

[0014] V i = x r1 + F(x r2 - x r3 )

[0015] i is the number of the individual on which the mutation operation is performed, V i is the intermediate individual obtained by the mutation operation, x r1 to x r3 are any three different individuals in the population to be optimized except the individual x i , and F is a preset scaling factor.

[0016] The crossover operation is represented by the following formula:

[0017]

[0018] x ij , V ij and U ij represent the individual x i , the corresponding intermediate individual V i and the mutated individual U iThe j-th decision variable, rand(0, 1) is a random number greater than 0 and less than 1, and cr is a preset crossover rate;

[0019] The selection operation includes, for any individual x i , if the objective function value of individual x i is greater than the objective function value of the corresponding mutated individual U i , replace individual x i with the mutated individual U i ;

[0020] Increment the optimization algebra by 1, and return to execute the steps of mutation, crossover, and selection operations for each individual in the population to be optimized in sequence until the optimization algebra is equal to the pattern threshold.

[0021] Optionally, the process of iteratively optimizing the unimproved individuals in the population to be optimized based on the artificial fish swarm algorithm includes:

[0022] Execute the optimization operation for each of the unimproved individuals;

[0023] Among them, the optimization operation includes:

[0024] Judge whether the clustering behavior of the unimproved individual is successful, and judge whether the chasing behavior of the unimproved individual is successful;

[0025] If only one of the clustering behavior and the chasing behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior;

[0026] If both the clustering behavior and the chasing behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller objective function value among the clustering behavior and the chasing behavior;

[0027] If both the clustering behavior and the chasing behavior of the unimproved individual are not successful, end the optimization operation;

[0028] After the optimization operations of all the unimproved individuals are completed, update the fish swarm parameters, and determine whether the optimization algebra is less than the preset termination threshold; increment the optimization algebra by 1, and return to execute the step of iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization algebra is equal to the termination threshold.

[0029] Optionally, judging whether the clustering behavior of the unimproved individual is successful includes:

[0030] For each unimproved individual x i , determine the neighborhood center x of the unimproved individual x i based on the following formulai，center :

[0031]

[0032] x i，L , x i，R , x i，U and x i，D are the von Neumann neighborhoods of the unimproved individual x i in the individual matrix; wherein, the individual matrix is composed of all individuals in the population to be optimized;

[0033] If the objective function value of the unimproved individual x i is greater than that of the neighborhood center x i，center , it is determined that the clustering behavior of the unimproved individual x i is successful;

[0034] If the objective function value of the unimproved individual x i is not greater than that of the neighborhood center x i，center , it is determined that the clustering behavior of the unimproved individual x i is not successful.

[0035] Optionally, determining whether the chasing behavior of the unimproved individual is successful includes:

[0036] For each unimproved individual x i , based on the following formula, determine the chasing individual x i corresponding to the unimproved individual x i，follow :

[0037] x i，follow = argmin(H(x i，L ), H(x i，R ), H(x i，U ), H(x i，D ))

[0038] H(x) represents a preset objective function. argmin represents selecting the individual that minimizes the value of the objective function H;

[0039] If the objective function value of the unimproved individual x i is greater than that of the chasing individual x i，follow , it is determined that the chasing behavior of the unimproved individual x i is successful;

[0040] If the objective function value of the unimproved individual x i is not greater than that of the chasing individual x i，follow , it is determined that the chasing behavior of the unimproved individual x i is not successful.

[0041] Optionally, the process of determining the decision variables for the decision lighting scheme includes:

[0042] Determine the light distribution information and the light fixture form information of the light fixtures according to the design conditions of the lighting scheme, and determine the type of light fixtures used according to the light distribution information and the light fixture form information of the light fixtures;

[0043] Under the constraints of the light distribution information and the light fixture form information of the light fixtures, determine the basic rules for light fixture layout, and extract the parameters for controlling the lighting scheme as alternative decision variables;

[0044] Determine the adjustable parameters among the alternative decision variables as decision variables.

[0045] The second aspect of this application provides a lighting scheme design device based on artificial fish swarm and differential evolution, including:

[0046] A determination unit for determining the reference value of the lighting design index and the pre-information of the target lighting space; wherein, the pre-information includes the size information, form information, light fixture layout range information and lighting work surface of the target lighting space;

[0047] An initialization unit for initializing the population to be optimized and the algorithm parameters; wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme;

[0048] A first optimization unit for iteratively optimizing the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset mode threshold;

[0049] A second optimization unit for performing mode switching and iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the number of optimization generations reaches a preset termination threshold; wherein, the mode switching is only performed once when the number of optimization generations reaches the mode threshold; the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population to be optimized; the unimproved individuals refer to the individuals that have not been updated when iteratively optimizing based on the differential evolution algorithm;

[0050] A selection unit is configured to select a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space; wherein, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of the individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables and variable values included in the individual and the pre-information of the target lighting space.

[0051] Optionally, when the first optimization unit performs iterative optimization on the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset pattern threshold, it is specifically configured to:

[0052] For each individual in the population to be optimized, perform mutation, crossover, and selection operations in sequence;

[0053] Among them, the mutation operation is represented by the following formula:

[0054] V i =x r1 +F(x r2 -x r3 )

[0055] i is the number of the individual on which the mutation operation is performed, V i is the intermediate individual obtained by the mutation operation, x r1 to x r3 are any three different individuals in the population to be optimized except the individual x i , and F is a preset scaling factor;

[0056] The crossover operation is represented by the following formula:

[0057]

[0058] x ij , V ij and U ij represent the j-th decision variable of the individual x i , the corresponding intermediate individual V i and the mutant individual U i in sequence, rand(0, 1) is a random number greater than 0 and less than 1, and cr is a preset crossover rate;

[0059] The selection operation includes, for any individual x i , if the objective function value of the individual x i is greater than the objective function value of the corresponding mutant individual U i , replace the individual x iReplace with mutant individual U i ;

[0060] Increment the optimization generation by 1, return and execute the mutation, crossover, and selection operation steps for each individual in the population to be optimized in sequence until the optimization generation is equal to the pattern threshold.

[0061] Optionally, when the second optimization unit iteratively optimizes the unimproved individuals in the population to be optimized based on the artificial fish swarm algorithm, it is specifically used for:

[0062] Execute an optimization operation for each unimproved individual;

[0063] Among them, the optimization operation includes:

[0064] Judge whether the clustering behavior of the unimproved individual is successful and whether the following behavior of the unimproved individual is successful;

[0065] If only one of the clustering behavior and the following behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior;

[0066] If both the clustering behavior and the following behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller objective function value among the clustering behavior and the following behavior;

[0067] If both the clustering behavior and the following behavior of the unimproved individual are not successful, end the optimization operation;

[0068] After the optimization operations of all unimproved individuals are completed, update the fish swarm parameters and determine whether the optimization generation is less than the preset termination threshold; increment the optimization generation by 1, return and execute the step of iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation is equal to the termination threshold.

[0069] The present application provides a method and device for designing an illumination scheme based on an artificial fish swarm and differential evolution. The method includes determining a reference value of illumination design indicators and pre-information of a target illumination space; wherein, the pre-information includes the size information, morphological information, lamp arrangement range information, and illumination working surface of the target illumination space; initializing a population to be optimized and algorithm parameters; wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining an illumination scheme; iteratively optimizing the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset mode threshold; iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the number of optimization generations reaches a preset termination threshold; wherein, the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population; unimproved individuals refer to the individuals that are not updated when iteratively optimizing based on the differential evolution algorithm; selecting a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the illumination scheme applicable to the target illumination space; wherein, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of an individual is determined according to the reference value and the individual value of the illumination design indicators corresponding to the individual; the individual value of the illumination design indicators corresponding to the individual is determined according to a set of decision variables and variable values included in the individual and the pre-information of the target illumination space. In the inverse design process of the illumination scheme, the present invention combines two algorithms, namely the differential evolution algorithm and the artificial fish swarm algorithm. Compared with the conventional method of using a single algorithm, it has a faster convergence speed and is not easily trapped in a local optimal solution. Therefore, based on the present invention, a lighting design scheme that meets the requirements can be obtained faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0071] Figure 1 It is a flowchart of a method for designing an illumination scheme based on an artificial fish swarm and differential evolution provided by an embodiment of the present application;

[0072] Figure 2 It is a schematic diagram of an individual matrix structure provided by an embodiment of the present application;

[0073] Figure 3 It is a flowchart of another method for designing an illumination scheme based on an artificial fish swarm and differential evolution provided by an embodiment of the present application;

[0074] Figure 4A Feu-illuminance-atmosphere relationship diagram provided by an embodiment of the present application;

[0075] Figure 5 A schematic diagram of the appearance of a lighting fixture provided by an embodiment of the present application;

[0076] Figure 6 A light distribution curve of a lighting fixture provided by an embodiment of the present application;

[0077] Figure 7 A schematic diagram of the correspondence between decision variables and the target lighting space provided by an embodiment of the present application;

[0078] Figure 8 A schematic diagram of the structure of a lighting scheme design device based on an artificial fish swarm and differential evolution provided by an embodiment of the present application. Detailed implementation manners

[0079] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0080] Combining the characteristics of the artificial fish swarm algorithm and the differential evolution algorithm, the present invention proposes a multi-strategy algorithm for inverse design (Inverse Design Multi-Strategy Algorithm, hereinafter referred to as IDMSA) for the inverse design of lighting schemes, and implements the lighting scheme design method described in the present invention based on the IDMSA algorithm. To facilitate the understanding of the technical solutions of the present application, some concepts and related configurations that may be involved in the present application are first introduced.

[0081] The target lighting space refers to a specific space for which a lighting scheme needs to be designed, and specifically may be one or more rooms in a certain building. Exemplarily, in a conference room where no lighting fixtures have been installed, several lighting fixtures need to be installed, and the layout method of these lighting fixtures is equivalent to the lighting scheme to be designed, and this conference room is equivalent to the target lighting space during design.

[0082] The pre-information of the target lighting space refers to some attributes of the target lighting space that need to be considered during the lighting scheme design, including but not limited to the function of the target lighting space, dimensional information (i.e., length, width, height, etc.), morphological information (cuboid, cylinder, or other irregular-shaped spaces), lighting fixture layout range information (used to describe the range within the target lighting space where lighting fixtures can be arranged), and lighting work surface (used to describe the position of the plane where visual work is carried out).

[0083] Generally, the pre - information of the target lighting space can be extracted from the architectural design model.

[0084] Lighting design indicators can be regarded as one or more indicators used to measure the quality of lighting design solutions. In different design methods, the quantity and type of lighting design indicators may vary. When selecting lighting design indicators, according to the lighting design concept and specific design conditions, the two major goals of "improving environmental quality and strengthening space function" under the connotation of lighting - building integration can be transformed into specific design indicators derived from luminance or illuminance, and further determine the reference values of the indicators required to achieve the design goals. And the goal of "optimizing lighting energy conservation" under the connotation of lighting - building integration is transformed into the minimization of lighting power density indicators.

[0085] In the present invention, the following three lighting design indicators are selected.

[0086] First, the work - plane illuminance. For the specific meaning and calculation method of the work - plane illuminance, reference can be made to the current relevant standards, such as the "Standard for Architectural Lighting Design", which will not be elaborated here.

[0087] Second, the space luminance coefficient Feu. The Feu index was proposed by Panasonic Corporation and is an index used to evaluate the subjective luminance perception of users with a single value. Feu achieves the unity of subjectivity and objectivity, has practicality, and is conducive to promotion. The definition formula of Feu is:

[0088] Feu = 1.5×Lg 0.7

[0089] In the above - mentioned definition formula, Lg is the geometric mean luminance in the target lighting space under a specific lighting scheme, which is numerically equal to the geometric mean of the luminance in the induced visual field, with the unit of candela per square meter (cd / m 2 ). In specific calculations, the part of the space where the luminance is greater than 1000 cd / m 2 is considered as the light source, which has little influence on the luminance perception and is not included in the geometric mean luminance. The induced visual field is the visual field range determined by the horizontal visual field of - 50° to 50° and the vertical visual field of - 50° to 35°. The specific calculation method of Lg can refer to the relevant existing technologies, which will not be elaborated here.

[0090] Third, the lighting power density (LPD), an indicator used to evaluate the optimization of lighting energy conservation, can be defined as the lighting installation power per unit area (including the power consumption of ballasts and transformers) of a room or place in a building, with the unit of watt per square meter.

[0091] It should be noted that among the above three lighting design indicators, reference values need to be set for the first two when designing the lighting scheme, while for the third one (LPD), only minimization is targeted in the design and no reference value needs to be set.

[0092] Decision variables, that is, a set of variables that can determine the lighting scheme. In other words, for a specific target lighting space, if the variable values of a set of decision variables can be determined, then based on these variable values, a set of lighting schemes applicable to the target lighting space can be uniquely determined, that is, the layout method of the luminaires within the target lighting space can be uniquely determined. In other words, a set of decision variables with a given set of variable values can be considered equivalent to a set of established lighting schemes.

[0093] When setting decision variables, first, according to the scheme design conditions, the light distribution information and luminaire form information of the luminaires need to be determined, and then the luminaire models to be used are selected based on the light distribution information and luminaire form information of the luminaires; then, under the limitation of the above information, the basic rules for luminaire layout are determined, and the parameters that control the specific layout scheme are extracted as alternative decision variables. For example, the alternative decision variables can include the luminous flux of each luminaire, the reflectivity of the wall / floor / ceiling, etc.; then, from these alternative decision variables, such as the luminous flux of each luminaire, the reflectivity of the wall / floor / ceiling, the adjustable parameters are determined as decision variables.

[0094] It should be noted that according to the differences in the target lighting space and related design requirements, the specific content of the decision variables is different, and this embodiment does not limit this.

[0095] As an example, for a cuboid shape, with the lighting working surface being a horizontal plane 750 mm above the ground and the lighting layout method being uniform grid lighting on the ceiling, in the present invention, the lateral spacing of the luminaires (the lateral direction refers to the width direction of the rectangular surface in the room) and the longitudinal spacing (the longitudinal direction refers to the depth direction of the rectangular surface), as well as the luminous flux of each luminaire, the reflectivity of the wall and the floor, can be selected as decision variables.

[0096] The objective function and constraint conditions are common concepts in various heuristic algorithms. When specifically implemented, heuristic algorithms generally aim to minimize (or maximize) the corresponding objective function value, and on the premise of controlling the optimized scheme to meet the constraint conditions, the given initial scheme is iteratively optimized to obtain a better scheme.

[0097] The objective function determined in the present invention is as follows:

[0098]

[0099] In the expression of the objective function, min H represents minimization optimization, that is, the optimization objective of the present invention is to minimize the function value of the objective function H, f iDenote the lighting design index for evaluating "environmental quality improvement and spatial function enhancement" of the i-th one. Combining the lighting design indices selected in the foregoing text, f in the above formula 1 can represent the working surface illuminance, and f 2 can represent Feu.

[0100] Denote the reference value of the i-th lighting design index. The specific value is set according to actual needs before implementing the method of the present invention. Combining the foregoing lighting design indices, can represent the reference value of the working surface illuminance, can represent the reference value of Feu.

[0101] LPD is the lighting power density, which evaluates the optimization of lighting energy conservation. C is a penalty coefficient, which is a large integer preset in advance. punish is the number of constraint conditions violated by the lighting scheme.

[0102] In the present invention, the constraint conditions include two requirements: whether it meets the illuminance uniformity and UGR standard value in the "Code for Architectural Lighting Design". That is to say, in the present invention, if the decision variable of a set of given variable values does not meet the illuminance uniformity requirement or does not meet the UGR standard value requirement, it is considered to violate one constraint condition. If it does not meet both the illuminance uniformity requirement and the UGR standard value requirement at the same time, it is considered to violate two constraint conditions.

[0103] UGR represents the Unified Glare Rating, which is a psychological parameter used to measure the subjective reaction of discomfort caused by the light emitted by lighting devices in the indoor visual environment to the human eye.

[0104] The value of the large integer C can be set according to requirements. Exemplarily, C can take the value of 100000 in the present invention.

[0105] The artificial fish swarm algorithm (AFSA), a widely used heuristic optimization algorithm, realizes optimization by constructing artificial fish individuals to imitate the foraging, aggregation and following behaviors of fish swarms.

[0106] The first aspect of the present application provides a lighting scheme design method based on artificial fish swarms and differential evolution, including:

[0107] S101, determine the reference value of the lighting design index and the pre-information of the target lighting space.

[0108] Among them, the pre-information includes the size information, morphological information, lamp layout range information and lighting working surface of the target lighting space.

[0109] The specific meanings of the lighting design indicators and the pre - information of the target lighting space have been described above.

[0110] The reference values of the lighting design indicators can be set according to the function of the target lighting space. For example, if the target lighting space is a meeting room for communication and discussion, the reference values can be set to meet the requirements of an open and active space atmosphere. If the target lighting space is a meeting room for making major decisions, the reference values can be designed to meet the requirements of a solemn space atmosphere.

[0111] S102, Initialize the population to be optimized and algorithm parameters.

[0112] Among them, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme.

[0113] The algorithm parameters that need to be initialized include: the total number of generations Ttotal (also known as the termination threshold), the population size N (used to limit the number of individuals in the population), the dimension n of the decision variables (used to limit the number of decision variables included in an individual), the upper and lower limit vectors Un and Ln of the decision variables (limiting the value range of the decision variable values), the step - size factor Step0 of the artificial fish - swarm algorithm (defined as the ratio of the initial step - size value to the width of the search domain) in the artificial fish - swarm algorithm, the step - size precision factor Step min0 (defined as the ratio of the minimum step - size Step min in the artificial fish - swarm algorithm to the initial step - size value), the bulletin board Board (used to record the optimal individual in each iterative optimization), the crossover rate cr of the differential evolution algorithm, and the scaling factor F.

[0114] Optionally, the process of determining the decision variables for the lighting scheme in step S102 may include:

[0115] According to the design conditions of the lighting scheme, determine the light - distribution information and lamp - form information of the lamps, and determine the lamp models used according to the light - distribution information and lamp - form information of the lamps;

[0116] Under the constraints of the light - distribution information and lamp - form information of the lamps, determine the basic rules for lamp layout, and extract the parameters controlling the lighting scheme as alternative decision variables;

[0117] Determine the adjustable parameters among the alternative decision variables as the decision variables.

[0118] In step S102, in addition to setting the specific values of the above - mentioned parameters that need to be initialized, the following operations can also be performed:

[0119] Set the current optimization generation T to 1, that is, T = 1.

[0120] Within the range of [Ln, Un], N individuals are randomly generated. Each individual contains a set of decision variables with given variable values. In other words, each individual corresponds to a lighting scheme. These N randomly generated individuals constitute the aforementioned population to be optimized.

[0121] Calculate the objective function value of each individual in the population to be optimized, and record the optimal individual on the bulletin board Board.

[0122] As described above, the expression of the objective function. The optimal individual refers to the individual with the smallest corresponding objective function value in the population to be optimized.

[0123] Calculate the following intermediate variables:

[0124] The search domain width ds is defined as the modulus of the diagonal vector of the search domain, and the calculation formula is as follows:

[0125] ds = ||Ln - Un||

[0126] The scaling factor S, when the upper and lower limits of the search domain are the upper and lower limits of the decision variables, is the corresponding length of the step size in the normalized search domain. The calculation formula is:

[0127]

[0128] The mode threshold Tm: a parameter that distinguishes the two stages of the IDMSA algorithm. When the number of optimization generations is less than this value, it is considered that the algorithm has not found the feasible domain under complex constraints. At this time, the mutation, crossover, and selection operations of the differential evolution algorithm are performed on the population to be optimized, that is, iterative optimization is performed based on the differential evolution algorithm to make as many individuals approach the feasible domain as quickly as possible; when the number of optimization generations is greater than this value, it is considered that most individuals have entered the feasible domain. At this time, the aggregation and following behaviors of the artificial fish swarm algorithm are performed to find the optimal solution within the feasible domain.

[0129] Optionally, Tm is set to 0.5Ttotal. That is to say, if Ttotal is set to 400, then Tm can be set to 200.

[0130] S103. Iteratively optimize the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches the preset mode threshold.

[0131] Optionally, the specific execution process of step S103 includes:

[0132] A1. Perform a mutation operation on each individual in the population to be optimized to obtain the corresponding intermediate individual for each individual.

[0133] Among them, the mutation operation includes calculating an intermediate individual based on three different individuals randomly selected from the population to be optimized except the individual on which the mutation operation is to be performed. The mutation operation can be specifically expressed by the following formula:

[0134] V i = x r1 + F(x r2 - x r3 )

[0135] In the formula of the above mutation operation, i represents the number of the individual on which the mutation operation is currently performed, that is, currently performing the mutation operation on the individual x i V i represents the intermediate individual obtained after the current mutation operation, and x r1 to x r3 are three different individuals randomly selected from the other individuals in the population to be optimized except the individual x i . As mentioned above, each individual contains a set of decision variables. When performing the mutation operation, the corresponding decision variables in the three selected individuals x r1 to x r3 are calculated according to the above formula to obtain new decision variables, and then the new decision variables are combined to obtain the corresponding intermediate individual.

[0136] A2. For each individual in the population to be optimized, perform the crossover operation to obtain the mutated individual of each individual.

[0137] Among them, the crossover operation includes mixing the decision variables in the individual on which the crossover operation is performed and the intermediate individual corresponding to the individual on which the crossover operation is performed to obtain the mutated individual.

[0138] The crossover operation can be specifically expressed by the following formula:

[0139]

[0140] In the formula of the above crossover operation, x ij represents the j-th decision variable of the individual x i on which the crossover operation is performed, V ij represents the j-th decision variable of the intermediate individual V i corresponding to the individual x i obtained in step A1, U ij represents the j-th decision variable of the mutated individual U i obtained by the crossover operation, rand(0, 1) represents a randomly generated number greater than 0 and less than 1, and cr is the crossover rate set during initialization.

[0141] The formula representation of the above crossover operation is as follows: for the j-th decision variable, a random number within the range of 0 to 1 is generated. If the random number is less than or equal to the crossover rate, the j-th decision variable of the mutant individual U i is determined to be the j-th decision variable of the intermediate individual V i , that is, U ij =V ij ; if the random number is greater than the crossover rate, the j-th decision variable of the mutant individual U i is determined to be the j-th decision variable of the individual x i , that is, U ij =x ij ; and so on until each decision variable of the mutant individual U i is determined.

[0142] A3. Perform the selection operation on each individual in the population to be optimized.

[0143] Among them, the selection operation includes comparing the first objective function value and the second objective function value. If the first objective function value is greater than the second objective function value, the individual to be subjected to the selection operation is replaced by the corresponding mutant individual. If the first objective function value is less than or equal to the second objective function value, the individual to be subjected to the selection operation is retained and the mutant individual corresponding to the individual to be subjected to the selection operation is removed; the first objective function value is the objective function value of the individual to be subjected to the selection operation, and the second objective function value is the objective function value of the mutant individual of the individual to be subjected to the selection operation.

[0144] Specifically, when performing the selection operation on the individual x i in the population, first calculate the objective function value corresponding to x i (i.e., the first objective function value) and the objective function value of the mutant individual U i . The specific calculation method can refer to the expression of the aforementioned objective function and will not be elaborated here.

[0145] Then, compare the first objective function value and the second objective function value. If the former is less than or equal to the latter, it means that the original individual x i is superior to the mutant individual U i . Retain the original individual in the population and delete the mutant individual U i ; if the former is greater than the latter, it means that the mutant individual U i is superior to the original individual x i . Replace the original individual x i with the mutant individual U i .

[0146] A4. After all the mutation, crossover, and selection operations on all individuals in the population are completed, determine whether the optimization generation is less than the pattern threshold.

[0147] If the optimization algebra is less than the mode threshold, execute step A5. If the optimization algebra is equal to the mode threshold, the iterative optimization based on the differential evolution algorithm alone terminates, and step S103 ends.

[0148] A5, increment the optimization algebra by 1.

[0149] After executing step A5, return to execute step A1, that is, perform mutation operations on each individual in the population to be optimized to obtain the intermediate individual steps corresponding to each individual until the optimization algebra is equal to the mode threshold.

[0150] Optionally, each time step A5 is executed, an optimal individual can also be selected from the current population to be optimized, and then the bulletin board is updated with this optimal individual.

[0151] S104, perform mode switching, and perform iterative optimization on the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization algebra reaches the preset termination threshold.

[0152] It should be noted that the mode switching step in step S104 is only executed once when the optimization algebra reaches the mode threshold, that is, when the optimization algebra is equal to the mode threshold, and is not executed subsequently; while the process of iterative optimization of individuals based on the differential evolution algorithm and the artificial fish swarm algorithm is repeatedly executed before the optimization algebra reaches the termination threshold.

[0153] Among them, an unimproved individual refers to an individual that has not been updated during iterative optimization based on the differential evolution algorithm.

[0154] When performing mode switching, the parameters required for subsequent execution of the artificial fish swarm algorithm can be set. Specifically, the initial value Step and the minimum step size Step of the step size when the artificial fish swarm algorithm is first executed can be set min of the initial value.

[0155] Where:

[0156] Step = Step0xds'

[0157] Step min = Step min0 ×Step

[0158] Among them, ds’ represents the search domain width of the current generation, and its specific calculation method is as follows:

[0159] First, calculate the decision variable range [Ln’, Un’] of the current generation:

[0160]

[0161] In the above formula, Denote the minimum value of the first - dimensional decision variable of all individuals in the current generation, and so on. Denote the minimum value of the n - dimensional decision variable of all individuals in the current generation; similarly, Denote the maximum value of the first - dimensional decision variable of all individuals in the current generation, Denote the maximum value of the n - dimensional decision variable of all individuals in the current generation;

[0162] Then, calculate the search - domain width of the current generation according to the decision - variable range of the current generation:

[0163] ds' = ||Ln' - Un'||

[0164] In step S104, each iterative optimization process includes:

[0165] First, optimize the individuals in the population to be optimized based on the differential evolution algorithm.

[0166] The specific execution process of the above - mentioned differential evolution algorithm is the same as that of the differential evolution algorithm in step S103, and will not be elaborated here.

[0167] Then, after the differential evolution algorithm in this iterative optimization ends, mark the individuals that have not been updated by the differential evolution algorithm in this iterative optimization as unimproved individuals.

[0168] Finally, optimize the above - marked unimproved individuals based on the artificial fish - swarm algorithm. After all unimproved individuals have executed the artificial fish - swarm algorithm, this iterative optimization ends and enters the next round of iterative optimization.

[0169] The following specifically introduces the meaning of unimproved individuals in combination with an example. For any individual x in the population to be optimized i , if in an iterative optimization process, when executing the differential evolution algorithm, individual x i is replaced by its corresponding mutant individual U i , then individual x i is marked as an improved individual. On the contrary, if in an iterative optimization process, when executing the differential evolution algorithm, individual x i is not replaced by its corresponding mutant individual U i , that is, after the differential evolution algorithm of this iterative optimization ends, the variable values of the decision variables in individual x i are the same as those before the differential evolution algorithm of this iterative optimization starts, then individual x i is considered an unimproved individual.

[0170] S105, Select a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space.

[0171] Among them, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of an individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables of the individual and the pre - information of the target lighting space.

[0172] According to the description of the decision variables in the previous text, after determining the optimal individual, a set of lighting schemes can be uniquely determined according to the set of decision variables containing given variable values in the optimal individual, and thus the design of the lighting scheme can be completed.

[0173] In each round of iterative optimization process of step S104, the specific execution process of optimizing the unimproved individuals with the artificial fish - swarm algorithm includes:

[0174] C1. Execute the optimization operation for each unimproved individual.

[0175] Among them, the optimization operation includes the following steps B1 to B4:

[0176] B1. Judge whether the clustering behavior of the unimproved individual is successful and whether the following - up behavior of the unimproved individual is successful.

[0177] First of all, in order to execute the artificial fish - swarm algorithm, all individuals in the population to be optimized need to be arranged into an Figure 2 individual matrix. In the matrix, each individual has four adjacent individuals: up, down, left, and right. For an individual x i , x i 's four adjacent individuals: up, down, left, and right form the von Neumann neighborhood of x i . Figure 2 Each circle in the figure represents an individual.

[0178] In particular, the individuals at both ends of a row are adjacent to each other. Similarly, the individuals at both ends of a column are adjacent to each other.

[0179] Based on the above settings, the process of judging whether the clustering behavior of the unimproved individual is successful in B1 includes:

[0180] Determine the von Neumann neighborhood of the unimproved individual in the individual matrix; among them, the individual matrix is composed of all individuals in the population to be optimized.

[0181] Taking the unimproved individual x i as an example, the von Neumann neighborhood of the unimproved individual x i is determined to be x i，L , x i，R , x i，U and x i，D .

[0182] Determine the neighborhood center of the von Neumann neighborhood of the unimproved individual.

[0183] The unimproved individual x i The neighborhood center x of i，center Can be calculated according to the following formula:

[0184]

[0185] Compare the third objective function value and the fourth objective function value.

[0186] Among them, the third objective function value is the objective function value of the unimproved individual; the fourth objective function value is the objective function value of the neighborhood center of the von Neumann neighborhood of the unimproved individual.

[0187] If the third objective function value is greater than the fourth objective function value, it means that the neighborhood center x of the unimproved individual i , center Is better than the unimproved individual x i , Determine that the clustering behavior of the unimproved individual is successful.

[0188] If the third objective function value is not greater than the fourth objective function value, it means that the unimproved individual x i Is better than the neighborhood center x of the unimproved individual i，center , Determine that the clustering behavior of the unimproved individual is not successful.

[0189] The process of judging whether the following behavior of the unimproved individual in B1 is successful includes:

[0190] Determine the von Neumann neighborhood of the unimproved individual in the individual matrix; among them, the individual matrix is composed of all individuals in the population to be optimized.

[0191] Taking the unimproved individual x i As an example, it is determined that the von Neumann neighborhood of the unimproved individual x i Is x i，L , x i，R , x i，U And x i，D .

[0192] Determine the optimal individual in the von Neumann neighborhood of the unimproved individual.

[0193] The optimal individual in the von Neumann neighborhood (which can also be denoted as the following individual, denoted by x i，follow Indicates) refers to the individual with the smallest corresponding objective function value in the von Neumann neighborhood. That is to say, for the unimproved individual x i , Its corresponding following individual is:

[0194] x i，follow = argmin(H(x i，L), H(x i，R ), H(x i，U ), H(x i，D ))

[0195] H(x) represents the objective function. argmin represents selecting the individual that minimizes the value of the objective function H.

[0196] Compare the third objective function value and the fifth objective function value.

[0197] Among them, the third objective function value is the objective function value of the unimproved individual; the fifth objective function value is the objective function value of the optimal individual in the von Neumann neighborhood of the unimproved individual.

[0198] If the third objective function value is greater than the fifth objective function value, it means that the unimproved individual x i corresponding tailgating individual x i，follow is superior to the unimproved individual x i , and it is determined that the tailgating behavior of the unimproved individual is successful.

[0199] If the third objective function value is not greater than the fifth objective function value, it means that the unimproved individual x i is superior to the unimproved individual x i corresponding tailgating individual x i，follow , and it is determined that the tailgating behavior of the unimproved individual is not successful.

[0200] B2. If only one of the clustering behavior and the tailgating behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior.

[0201] B3. If both the clustering behavior and the tailgating behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller corresponding objective function value among the clustering behavior and the tailgating behavior.

[0202] Specifically, in steps B2 and B3, the update method for the unimproved individual x i is as follows:

[0203] Determine the target position xt. For step B2, if the successful behavior is the clustering behavior, then determine the target position as the aforementioned central individual x i，center , that is, xt = x i，center , if the successful behavior is the tailgating behavior, then determine the target position as the aforementioned tailgating individual x i，follow , that is, xt = x i，follow ; for step B3, the target position can be determined as the individual with the smallest corresponding objective function value among the central individual and the tailgating individual.

[0204] Perform search domain normalization and convert the unimproved individual x i and the target position into the normalized individual x i’ and the normalized position xt’, the conversion formula is as follows:

[0205]

[0206] In the above conversion formula, the fraction means dividing each element of the numerator by the corresponding element in the denominator.

[0207] Update the normalized individual in the normalized space to obtain the updated normalized individual x i，new ’, and the update formula is as follows:

[0208]

[0209] Restore the updated normalized individual to the individual in the original search domain to obtain the updated unimproved individual x i，new , and the restoration formula is as follows:

[0210] x i，new = Ln + (Un - Ln) × x i，new ′

[0211] In the restoration formula, (Un - Ln) × x i，new ’ means multiplying each decision variable of Un - Ln by the corresponding normalized value in the updated normalized individual x i，new ’.

[0212] Finally, replace the original unimproved individual with the updated unimproved individual to complete this update.

[0213] Specifically, if the value of a certain decision variable in the updated unimproved individual exceeds the value range determined by [Ln, Un], the value of this decision variable will be replaced with the closer boundary value. That is to say, if the value of a certain decision variable is greater than the upper limit of the value range, it will be replaced with the upper limit; if it is less than the lower limit, it will be replaced with the lower limit.

[0214] B4, if both the clustering behavior and the chasing behavior of the unimproved individual are not successful, end the optimization operation.

[0215] C2, after completing the optimization operation of each unimproved individual, update the fish swarm parameters and determine whether the optimization generation is less than the preset termination threshold.

[0216] The above update is mainly to adaptively update the step size Step, and specifically, it can be carried out according to the following formula:

[0217]

[0218] In the above formula, Step T-1 represents the step size before update, Step T represents the step size after update, Step minIt represents the aforementioned minimum step size, and T is the current optimization generation number.

[0219] C3. If the optimization generation number is less than the termination threshold, increment the optimization generation number by 1, and return to execute the step of optimizing the individuals in the population to be optimized based on the differential evolution algorithm in step S104 until the optimization generation number is equal to the termination threshold.

[0220] Optionally, in step C3, the optimal individual can also be selected from the current population to be optimized, and then the bulletin board is updated with this optimal individual.

[0221] This application provides a method and device for designing a lighting scheme based on artificial fish swarm and differential evolution. The method includes determining the reference value of the lighting design index and the pre-information of the target lighting space; where the pre-information includes the size information, shape information, lamp layout range information, and lighting work surface of the target lighting space; initializing the population to be optimized and algorithm parameters; where the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme; iteratively optimizing the population to be optimized based on the differential evolution algorithm until the optimization generation number reaches the preset mode threshold; iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation number reaches the preset termination threshold; where the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population; the unimproved individuals refer to the individuals that are not updated when iteratively optimizing based on the differential evolution algorithm; selecting a set of decision variables and their variable values included in the optimal individual in the population to be optimized as the decision variables and their variable values of the lighting scheme applicable to the target lighting space; where the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of an individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables of the individual and the pre-information of the target lighting space. In the inverse design process of the lighting scheme, the present invention combines two algorithms, the differential evolution and the artificial fish swarm algorithm. Compared with the conventional method using a single algorithm, it has a faster convergence speed and is not easily trapped in a local optimal solution. Therefore, based on the present invention, a lighting design scheme that meets the requirements can be obtained faster.

[0222] Please refer to Figure 3 which is the flowchart of another method for designing a lighting scheme based on artificial fish swarm and differential evolution provided by the embodiments of this application.

[0223] S301, Initialization.

[0224] Step S301 is equivalent to step S102 of the foregoing embodiment.

[0225] S302, Differential evolution operation.

[0226] S303. Compare T and Tm.

[0227] If T is less than Tm, execute step S311; if T is equal to Tm, execute step S304; if T is greater than Tm, execute step S305. Here, Tm is the aforementioned pattern threshold.

[0228] The loop of steps S302, S303, S310, S311, and S312 is equivalent to step S103 of the aforementioned embodiment.

[0229] S304. Perform mode switching.

[0230] After step S304 is executed, execute step S305.

[0231] S305. Cluster and pursue the unimproved individuals.

[0232] S306. Is there a successful result?

[0233] Steps S305 and S306 are equivalent to step B1 of the aforementioned embodiment.

[0234] If there is a successful result, execute step S307; if there is no successful result, execute step S310.

[0235] S307. Are both actions successful?

[0236] If only one of the two actions is successful, execute step S308; if both actions are successful, execute step S309.

[0237] S308. Execute the successful result.

[0238] S309. Execute the better result.

[0239] Steps S307 to S309 are equivalent to steps B2 to B4 of the aforementioned embodiment.

[0240] Steps S305 to S309 are equivalent to step C1 of the aforementioned embodiment.

[0241] S310. Update the fish swarm parameters.

[0242] Step S310 is equivalent to step C2 of the aforementioned embodiment.

[0243] S311. Update the bulletin board.

[0244] S312. Is T < Ttotal?

[0245] If T is less than Ttotal, execute step S313; if T is not less than Ttotal, execute step S314.

[0246] Among them, Ttotal is the aforementioned termination threshold.

[0247] S313, T = T + 1.

[0248] After the execution of step S313 is completed, return to execute step S302.

[0249] Steps S312 to S313 are equivalent to step C3 of the aforementioned embodiment.

[0250] Steps S302 to S312 are equivalent to step S104 of the aforementioned embodiment; among them, it can be seen that the mode switching described in S304 is only executed once when the optimization algebra T is equal to the mode threshold Tm, and is not executed in other cases.

[0251] S314, output the result.

[0252] Step S314 is equivalent to step S105 of the aforementioned embodiment.

[0253] To facilitate the understanding of the lighting scheme design method based on artificial fish swarm and differential evolution provided by this application, the execution process of this method will be described below with specific examples.

[0254] In particular, to highlight the beneficial effects of the present invention, the results of using other methods to design the lighting scheme for this example are further provided for comparison.

[0255] Considering various types of public spaces, the meeting room space is more important for the production activities of urban workers. Users spend more time engaged in visual work in the meeting room compared to other example types. Therefore, higher lighting quality, a spatial atmosphere corresponding to the activities, and energy conservation as much as possible are more needed. The requirements for lighting quality in the meeting room space are: 1. Provide sufficient lighting to maintain visual work; 2. Ensure the comfort of visual work and avoid eye fatigue. The demand for the lighting spatial atmosphere in the meeting room is more diverse: for design work, the meeting room is often used for communication and discussion, which requires a more active spatial atmosphere; for administrative work, the meeting room is used to decide major collective matters, which requires a more solemn spatial atmosphere.

[0256] In this example, the IDMSA algorithm is used to perform lighting inverse design on a certain meeting room space. The main function of the target lighting space is communication and discussion. Therefore, a lighting scheme that meets the two conditions of feeling brighter and having an open atmosphere needs to be designed.

[0257] Panasonic Corporation proposed that the atmosphere of the lighting space is jointly shaped by Feu and the work surface illuminance. High illuminance + high Feu shapes an open atmosphere, low illuminance + high Feu shapes a soft atmosphere; high illuminance + low Feu shapes a serene atmosphere; low illuminance + low Feu shapes a relaxed atmosphere. The Feu-illuminance-atmosphere relationship in the meeting room space is shown inFigure 4 。

[0258] In addition, the research of Panasonic Corporation also gives the corresponding relationship between the Feu value of the conference room space and the brightness perception: when Feu is between 6 and 10, it feels relatively dim; when it is between 10 and 16, the brightness feeling is moderate; when it is between 16 and 24, it feels relatively bright.

[0259] According to Figure 4 the Feu-illuminance-atmosphere relationship in

[0260] In addition to IDMSA, two algorithms, namely the Differential Evolution (hereinafter referred to as DE) and the Artificial Fish Swarm Algorithm (hereinafter referred to as AFSA), are also used to design this example.

[0261] In order to compare the advantages and disadvantages of different algorithms, several indicators for measuring the quality of algorithms, that is, the indicators for performance comparison, are provided below, including:

[0262] (1) Failure rate: Based on a specific algorithm, the inverse design of the conference room lighting scheme is repeatedly executed under the same conditions for multiple times. Each time, an optimal individual and the corresponding objective function value can be obtained. The acceptable range is set as the objective function value ≤ 10. If the corresponding objective function value does not reach the range, it is regarded as a failure in one run. The failure rate is the ratio of the number of failures to the number of repetitions.

[0263] Because lighting simulation takes a long time and the cost of re-optimization after optimization failure is too high, this indicator is the primary evaluation criterion for lighting optimization problems.

[0264] (2) Best solution: As mentioned above, based on a specific algorithm, an optimal individual and the corresponding objective function value can be obtained for each repeated experiment. The best solution is the optimal (the smallest in this invention) one among the multiple objective function values obtained from multiple repeated experiments.

[0265] This indicator examines whether the algorithm can find the global optimal solution of the function to be optimized.

[0266] (3) Average solution for each time: As mentioned above, the average solution for each time is the average value of the above-mentioned multiple objective function values.

[0267] This indicator is used to evaluate the optimization ability of the algorithm as a whole.

[0268] (4) Standard deviation: As mentioned above, the standard deviation is the value of the standard deviation of the above-mentioned multiple objective function values.

[0269] This indicator examines the stability of the algorithm.

[0270] (5) Average duration: The average value of the time consumed for each execution when a specific algorithm repeatedly executes the inverse design of the conference room lighting scheme.

[0271] This metric can estimate the efficiency of the algorithm. However, since the duration is related to the algorithm program writing, it is for reference only.

[0272] When evaluating the best solution, each average solution, standard deviation, and average duration, the data of failed runs are excluded. In the present invention, the smaller the above metric value, the better the algorithm performance.

[0273] As an example, the parameter configurations of each algorithm used in this example can adopt the configurations in Table 1 below:

[0274] Table 1

[0275]

[0276]

[0277] The specific implementation steps of this example are as follows:

[0278] First, determine the pre-information of the target lighting space, that is, the scale, the range where lights can be installed, and determine the lighting index reference values according to the building space function. In this example, the width of the conference room space is 6000 mm, the depth is 9000 mm, and the clear height under the ceiling is 2970 mm. For aesthetic considerations, no lights are installed within 500 mm around the ceiling. According to the aforementioned atmosphere requirements and Figure 4 the shown correspondence, the Feu reference value is taken as 20, and the illuminance reference value is taken as 500 lx.

[0279] Secondly, determine the decision variables. In this conference room space, the adjustable decision variables include the lateral (width direction) spacing, longitudinal (depth direction) spacing of the lights, the luminous flux of each light, the reflectivity of the wall and floor.

[0280] The ceiling material has been determined, and the reflectivity is 0.7. The parameters of the lights are shown in Table 2. The appearance and light distribution curve of the lights can be referred to Figure 5 and Figure 6 .

[0281] Table 2

[0282]

[0283] The units and value ranges of the above decision variables are shown in Table 3:

[0284] Table 3 List of decision variables for the inverse design of conference room space lighting

[0285]

[0286]

[0287] According to Figure 7 the schematic diagram of the correspondence between the decision variables and the target illumination space shown, a set of decision variables with given variable values can be converted into corresponding illumination schemes (lamp layout schemes). Specifically, d1 and d2 in the figure correspond to the horizontal spacing and vertical spacing in the decision variables in sequence. Taking the face width direction as the positive x-axis direction, the depth direction as the positive y-axis direction, and the corner point of the room as the origin to establish a plane rectangular coordinate system, and using the center point of the lamp as the lamp layout coordinate, the process of converting the decision variables into the lamp layout scheme can be described as follows:

[0288] 1) Select the area within 500 mm from the ceiling boundary as the lamp layout area ( Figure 7 the area within the dashed box in

[0289] 2) Calculate the number of horizontally arranged lamps xn as:

[0290]

[0291] That is, the maximum number of lamps that can be arranged when the horizontal spacing is d1. Let the number of horizontally arranged lamps be xn, then the abscissa of the lamp layout grid is calculated according to the following formula:

[0292]

[0293] where i = 0, 1, 2…, xn - 1.

[0294] 3) The number of vertically arranged lamps and the ordinate of the lamp layout grid can be obtained by referring to step 2) and using the vertical spacing and the depth for calculation, which will not be elaborated here.

[0295] 4) Arrange lamps on the ceiling according to the abscissa and ordinate of the grid intersection points, and the luminous flux of each lamp has been specified by the decision variables.

[0296] Again, set the objective function and constraint conditions. According to the lighting design requirements, referring to the expression of the objective function and the lighting design indicators mentioned above, the objective function is set in this example as:

[0297]

[0298] For the definitions of each parameter in the above formula, please refer to the introduction of the objective function concept and the corresponding expression mentioned above.

[0299] There are two constraint conditions set in this example, namely, the illuminance uniformity ≥ 0.6 and UGR ≤ 19.

[0300] Since the luminous flux of the lamps in this example can be arbitrarily variable and it is not suitable to calculate LPD using the rated power of the lamps, the calculation formula of LPD in this example is set as:

[0301]

[0302] In the formula, Phi is the total luminous flux of all lamps, A is the room area, and 100 represents that the luminous efficacy of the used LED lamps is 100 lm / W.

[0303] Finally, run three algorithms, namely IDMSA, DE, and AFSA, according to the parameters shown in Table 1, and record the optimal individuals given by each algorithm.

[0304] The values of the performance evaluation indicators output after the operation of IDMSA, DE, and AFSA can be seen in Table 4.

[0305] Table 4

[0306]

[0307] What Table 4 reflects is the energy consumption of the scheme on the premise of meeting the requirements of Feu and illuminance indicators. For the best solution, the sorting of the three algorithms is IDMSA, AFSA, and DE in turn, that is, the objective function value corresponding to a set of decision variables included in the optimal individual output by the IDMSA algorithm is the smallest, and the sorting of the average solutions each time is the same as the best result; the sorting of stability (the smaller the standard deviation, the better the stability) and duration is IDMSA, DE, and AFSA in turn. To sum up, IDMSA shows better performance than the individual DE and AFSA in all indicators, indicating that the improvement is effective.

[0308] To verify that IDMSA has indeed achieved the improvement effect, an independent variable t-test is used to judge whether the better performance shown by IDMSA is due to random factors. Table 5 shows the comparison results of the average solutions obtained by IDMSA in 10 runs and the average solutions of AFSA and DE. The significance indicators of the two algorithms AFSA and DE are both less than 0.05, and the significance indicator of DE is less than 0.01, indicating that the differences between the average solutions of IDMSA and the other two algorithms are all statistically significant, further proving that the IDMSA algorithm proposed in the present invention combines the advantages of DE and AFSA, realizes the improvement of operation efficiency and search accuracy, and is suitable for solving various types of lighting inverse design problems.

[0309] Table 5, Independent variable t-test results between the average solutions output by the IDMSA algorithm and AFSA and DE

[0310]

[0311]

[0312] As can be seen from the above embodiments and corresponding examples, the lighting scheme design method based on artificial fish swarm and differential evolution provided by the present invention can generate the most energy-efficient lighting design scheme on the premise of meeting the brightness and illuminance conditions, and realizes the design goals of "improving environmental quality, strengthening space function, and optimizing lighting energy conservation" for lighting building integration by means of intelligence; aiming at the "black box" characteristic of lighting inverse design optimization, the present invention organically combines two heuristic algorithms, namely the artificial fish swarm algorithm and the differential evolution algorithm, controls the switching of the operation modes of the two algorithms through the new parameter "mode threshold", executes the clustering and chasing behaviors of the artificial fish swarm algorithm based on the von Neumann neighborhood, and applies an adaptive strategy to the step size parameter of the artificial fish swarm algorithm, solves the problems existing in the existing heuristic optimization algorithms such as slow convergence speed and falling into local optimal solutions, and realizes efficient and accurate integrated lighting design.

[0313] According to the lighting scheme design method based on artificial fish swarm and differential evolution provided by the embodiments of the present application, the embodiments of the present application also provide a lighting scheme design device based on artificial fish swarm and differential evolution. Please refer to Figure 8 , and the device may include the following units.

[0314] A determination unit 801, configured to determine a reference value of lighting design indicators and pre-information of a target lighting space.

[0315] Wherein, the pre-information includes size information, morphological information, lamp layout range information, and lighting work surface of the target lighting space.

[0316] An initialization unit 802, configured to initialize a population to be optimized and algorithm parameters.

[0317] Wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining a lighting scheme.

[0318] A first optimization unit 803, configured to perform iterative optimization on the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches a preset mode threshold.

[0319] A second optimization unit 804, configured to perform mode switching, and perform iterative optimization on the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the number of optimization generations reaches a preset termination threshold.

[0320] It should be noted that the second optimization unit 804 only performs mode switching once when the number of optimization generations reaches (i.e., equals) the mode threshold, and does not perform it subsequently. However, the iterative optimization of the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm is repeatedly performed before the number of optimization generations reaches the termination threshold.

[0321] Among them, the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population to be optimized; the unimproved individuals refer to the individuals that have not been updated during the iterative optimization based on the differential evolution algorithm.

[0322] The selection unit 805 is configured to select a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space.

[0323] Among them, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of an individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to an individual is determined according to a set of decision variables and variable values of the individual and the pre-information of the target lighting space.

[0324] Optionally, when the first optimization unit 803 performs iterative optimization on the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches the preset mode threshold, it is specifically configured to:

[0325] Perform a mutation operation on each individual in the population to be optimized to obtain an intermediate individual corresponding to each individual; among them, the mutation operation includes calculating the intermediate individual according to three individuals randomly selected from the population to be optimized except the individual on which the mutation operation is performed.

[0326] Perform a crossover operation on each individual in the population to be optimized to obtain a mutant individual corresponding to each individual; among them, the crossover operation includes mixing the decision variables of the individual on which the crossover operation is performed and the intermediate individual corresponding to the individual on which the crossover operation is performed to obtain the mutant individual.

[0327] Perform a selection operation on each individual in the population to be optimized; among them, the selection operation includes comparing the first objective function value and the second objective function value. If the first objective function value is greater than the second objective function value, replace the individual on which the selection operation is performed with the corresponding mutant individual. If the first objective function value is less than or equal to the second objective function value, retain the individual on which the selection operation is performed and remove the mutant individual corresponding to the individual on which the selection operation is performed; the first objective function value is the objective function value of the individual on which the selection operation is performed, and the second objective function value is the objective function value of the mutant individual of the individual on which the selection operation is performed.

[0328] If the number of optimization generations is less than the preset mode threshold and less than the preset termination threshold, increment the number of optimization generations by 1, and return to execute the step of performing a mutation operation on each individual in the population to be optimized to obtain an intermediate individual corresponding to each individual until the number of optimization generations is equal to the mode threshold.

[0329] Optionally, when the second optimization unit 804 performs iterative optimization on the unimproved individuals in the population to be optimized based on the artificial fish swarm algorithm, it is specifically used for:

[0330] Performing an optimization operation on each unimproved individual;

[0331] Among them, the optimization operation includes:

[0332] Judging whether the clustering behavior of the unimproved individual is successful and judging whether the chasing behavior of the unimproved individual is successful;

[0333] If only one of the clustering behavior and the chasing behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior;

[0334] If both the clustering behavior and the chasing behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller corresponding objective function value among the clustering behavior and the chasing behavior;

[0335] If both the clustering behavior and the chasing behavior of the unimproved individual are not successful, end the optimization operation;

[0336] After the optimization operations of all unimproved individuals are completed, update the fish swarm parameters and determine whether the optimization generation is less than the preset termination threshold; if the optimization generation is less than the termination threshold, increment the optimization generation by 1, and return to execute the step of iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation is equal to the termination threshold.

[0337] Optionally, when the second optimization unit 804 judges whether the clustering behavior of the unimproved individual is successful, it is specifically used for:

[0338] Determining the von Neumann neighborhood of the unimproved individual in the individual matrix; where the individual matrix is composed of all unimproved individuals;

[0339] Determining the neighborhood center of the von Neumann neighborhood of the unimproved individual;

[0340] Comparing the third objective function value and the fourth objective function value; where the third objective function value is the objective function value of the unimproved individual; the fourth objective function value is the objective function value of the neighborhood center of the von Neumann neighborhood of the unimproved individual;

[0341] If the third objective function value is greater than the fourth objective function value, it is determined that the clustering behavior of the unimproved individual is successful;

[0342] If the third objective function value is not greater than the fourth objective function value, it is determined that the clustering behavior of the unimproved individual is not successful.

[0343] Optionally, when the second optimization unit 804 judges whether the chasing behavior of the unimproved individual is successful, it is specifically used for:

[0344] Determine the von Neumann neighborhood of unimproved individuals in the individual matrix; wherein, the individual matrix consists of all unimproved individuals;

[0345] Determine the optimal individual in the von Neumann neighborhood of the unimproved individual;

[0346] Compare the third objective function value and the fifth objective function value; wherein, the third objective function value is the objective function value of the unimproved individual; the fifth objective function value is the objective function value of the optimal individual in the von Neumann neighborhood of the unimproved individual;

[0347] If the third objective function value is greater than the fifth objective function value, determine that the chasing behavior of the unimproved individual is successful;

[0348] If the third objective function value is not greater than the fifth objective function value, determine that the chasing behavior of the unimproved individual is unsuccessful.

[0349] For the device provided in this embodiment, its specific working principle can refer to the relevant steps in the lighting scheme design method based on artificial fish swarm and differential evolution provided in any embodiment of this application, which will not be elaborated here.

[0350] This application provides a lighting scheme design device based on artificial fish swarm and differential evolution, including: a determination unit 801 determines the reference value of lighting design indicators and the pre-information of the target lighting space; an initialization unit 802 initializes the population to be optimized and algorithm parameters; each individual in the population to be optimized includes a set of decision variables for determining the lighting scheme; a first optimization unit 803 and a second optimization unit 804 iteratively optimize the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm in sequence until the number of optimization generations reaches the termination threshold, and finally a selection unit 805 selects a set of decision variables and their variable values included in the optimal individual in the population to be optimized as the decision variables and their variable values of the lighting scheme applicable to the target lighting space. In the inverse design process of the lighting scheme, this invention combines two algorithms, namely the differential evolution algorithm and the artificial fish swarm algorithm. Compared with the conventional method of using a single algorithm, it has a faster convergence speed and is not easily trapped in a local optimal solution. Therefore, based on this invention, a lighting design scheme that meets the requirements can be obtained faster.

[0351] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising said element.

[0352] It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to define the order or interdependence of the functions performed by these devices, modules or units.

[0353] Those skilled in the art can implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A lighting scheme design method based on artificial fish swarm and differential evolution, characterized in that, Including: Determine the reference value of the lighting design index and the pre-information of the target lighting space; wherein, the pre-information includes the size information, morphological information, lamp layout range information and lighting work surface of the target lighting space; Initialize the population to be optimized and algorithm parameters; wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme; Iteratively optimize the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches the preset mode threshold; Perform mode switching, and iteratively optimize the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the number of optimization generations reaches the preset termination threshold; wherein, the mode switching is only performed once when the number of optimization generations reaches the mode threshold; the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population to be optimized; the unimproved individuals refer to the individuals that have not been updated during the iterative optimization based on the differential evolution algorithm; Select a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space; wherein, the optimal individual is the individual with the smallest corresponding objective function value among all individuals in the population to be optimized; the objective function value of the individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables and variable values included in the individual and the pre-information of the target lighting space; Wherein, the expression of the objective function is: min H represents minimizing the optimization, f i represents the lighting design index for the i-th evaluation of "environmental quality improvement and spatial function enhancement", represents the reference value of the i-th lighting design index, LPD is the lighting power density, C is a preset penalty coefficient, and punish represents the number of lighting scheme violations of the constraints.

2. The method according to claim 1, wherein The process of iteratively optimizing the population to be optimized based on the differential evolution algorithm until the number of optimization generations reaches the preset mode threshold includes: For each individual in the population to be optimized, perform mutation, crossover and selection operations in sequence; Wherein, the mutation operation is represented by the following formula: V i = x r1 + F(x r2 - x r3 ) i is the number of the individual to be mutated, V i is the intermediate individual obtained by the mutation operation, x r1 to x r3 are any three different individuals in the population to be optimized except the individual x i ; F is a preset scaling factor. The crossover operation is represented by the following formula: x ij , V ij and U ij respectively represent the j-th decision variable of the individual x i , the corresponding intermediate individual V i and the mutant individual U i , where rand(0, 1) is a random number greater than 0 and less than 1, and cr is a preset crossover rate; The selection operation includes, for any individual x i , if the objective function value of individual x i is greater than the objective function value of the corresponding mutated individual U i , replace individual x i with the mutated individual U i ; Increment the number of optimization generations by 1, and return to execute the step of performing mutation, crossover and selection operations in sequence for each individual in the population to be optimized until the number of optimization generations is equal to the mode threshold; 3. The method according to claim 1, characterized in that The process of iteratively optimizing the unimproved individuals in the population to be optimized based on the artificial fish swarm algorithm includes: Perform an optimization operation for each unimproved individual; Wherein, the optimization operation includes: Judge whether the aggregation behavior of the unimproved individual is successful, and judge whether the following behavior of the unimproved individual is successful; If only one of the aggregation behavior and the following behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior; If both the aggregation behavior and the following behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller corresponding objective function value among the aggregation behavior and the following behavior; If both the aggregation behavior and the following behavior of the unimproved individual are not successful, end the optimization operation; After all the optimization operations on the unimproved individuals are completed, update the fish swarm parameters and determine whether the optimization generation is less than the preset termination threshold; increment the optimization generation by 1, and return to execute the step of iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation is equal to the termination threshold.

4. The method according to claim 3, characterized in that, The determination of whether the clustering behavior of the unimproved individuals is successful includes: For each of the unimproved individuals x i , determine the neighborhood center x i of the unimproved individual x i，center according to the following formula: x i，L , x i，R , x i，U and x i，D are the von Neumann neighborhoods of the unimproved individual x i in the individual matrix; wherein, the individual matrix is composed of all individuals in the population to be optimized; If the objective function value of the unimproved individual x i is greater than that of the neighborhood center x i，center , determine that the clustering behavior of the unimproved individual x i is successful; If the objective function value of the unimproved individual x i is not greater than that of the neighborhood center x i，center , determine that the clustering behavior of the unimproved individual x i has not been successful.

5. The method according to claim 3, characterized in that, The determination of whether the chasing behavior of the unimproved individuals is successful includes: For each of the said unimproved individuals x i , determine the corresponding rear-ended individual x i of the unimproved individual x based on the following formula i，follow : x i，follow = argmin(H(x i，L ), H(x i，R ), H(x i，U ), H(x i，D )) H(x) represents a preset objective function, and argmin represents selecting the individual that minimizes the value of the objective function H; If the objective function value of the unimproved individual x i is greater than the objective function value of the chased individual x i，follow , it is determined that the chasing behavior of the unimproved individual x i is successful; If the objective function value of the unimproved individual x i is not greater than that of the chased individual x i，follow , it is determined that the chasing behavior of the unimproved individual x i has not been successful.

6. The method according to any one of claims 1 to 5, characterized in that The process of determining the decision variables for the lighting scheme includes: According to the design conditions of the lighting scheme, determine the light distribution information and the lamp form information of the lamps, and determine the lamp models used according to the light distribution information and the lamp form information of the lamps; Under the constraints of the light distribution information and the lamp form information of the lamps, determine the basic rules for lamp layout, and extract the parameters for controlling the lighting scheme as alternative decision variables; Determine the parameters that can be adjusted among the alternative decision variables as the decision variables.

7. An illumination scheme design device based on an artificial fish swarm and differential evolution, characterized in that, Including: A determination unit for determining the reference value of the lighting design index and the pre-information of the target lighting space; wherein, the pre-information includes the size information, form information, lamp layout range information and lighting work surface of the target lighting space; An initialization unit for initializing the population to be optimized and the algorithm parameters; wherein, the population to be optimized includes multiple individuals, and each individual includes a set of decision variables for determining the lighting scheme; A first optimization unit for iteratively optimizing the population to be optimized based on the differential evolution algorithm until the optimization generation reaches the preset mode threshold; A second optimization unit for performing mode switching and iteratively optimizing the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation reaches the preset termination threshold; wherein, the mode switching is only performed once when the optimization generation reaches the mode threshold; the artificial fish swarm algorithm is used to optimize the unimproved individuals in the population to be optimized; the unimproved individuals refer to the individuals that are not updated during the iterative optimization based on the differential evolution algorithm; A selection unit for selecting a set of decision variables and variable values included in the optimal individual in the population to be optimized as the decision variables and variable values of the lighting scheme applicable to the target lighting space; wherein, the optimal individual is the individual with the smallest corresponding objective function value among all the individuals in the population to be optimized; the objective function value of the individual is determined according to the reference value and the individual value of the lighting design index corresponding to the individual; the individual value of the lighting design index corresponding to the individual is determined according to a set of decision variables and variable values included in the individual and the pre-information of the target lighting space; Wherein, the expression of the objective function is: min H represents minimizing the optimization, f i represents the i-th lighting design index for evaluating "environmental quality improvement and spatial function enhancement", represents the reference value of the i-th lighting design index, LPD is the lighting power density, C is a preset penalty coefficient, and punish represents the number of lighting schemes that violate the constraint conditions.

8. The device according to claim 7, wherein When the first optimization unit iteratively optimizes the population to be optimized based on the differential evolution algorithm until the optimization generation reaches the preset mode threshold, it specifically is used for: For each individual in the population to be optimized, perform mutation, crossover, and selection operations in sequence; Among them, the mutation operation is represented by the following formula: V i = x r1 + F(x r2 - x r3 ) i is the number of the individual to be mutated, V i is the intermediate individual obtained by the mutation operation, x r1 to x r3 are any three different individuals in the population to be optimized except the individual x i ; F is a preset scaling factor. The crossover operation is represented by the following formula: x ij , V ij and U ij successively represent the j-th decision variable of the individual x i , the corresponding intermediate individual V i and the mutant individual U i , where rand(0, 1) is a random number greater than 0 and less than 1, and cr is a preset crossover rate; The selection operation includes, for any individual x i , if the objective function value of individual x i is greater than the objective function value of the corresponding mutated individual U i , replace individual x i with the mutated individual U i ; Increment the optimization generation by 1, and return to execute the step of performing mutation, crossover, and selection operations in sequence for each individual in the population to be optimized until the optimization generation is equal to the pattern threshold.

9. The device according to claim 7, characterized in that, When the second optimization unit performs iterative optimization on the unimproved individuals in the population to be optimized based on the artificial fish swarm algorithm, it is specifically used for: Perform optimization operations on each of the unimproved individuals; Among them, the optimization operations include: Judge whether the clustering behavior of the unimproved individual is successful, and judge whether the following behavior of the unimproved individual is successful; If only one of the clustering behavior and the following behavior of the unimproved individual is successful, update the unimproved individual according to the successful behavior; If both the clustering behavior and the following behavior of the unimproved individual are successful, update the unimproved individual according to the behavior with the smaller objective function value among the clustering behavior and the following behavior; If both the clustering behavior and the following behavior of the unimproved individual are not successful, end the optimization operation; After the optimization operations of all the unimproved individuals are completed, update the fish swarm parameters, and determine whether the optimization generation is less than the preset termination threshold; increment the optimization generation by 1, and return to execute the step of performing iterative optimization on the individuals in the population to be optimized based on the differential evolution algorithm and the artificial fish swarm algorithm until the optimization generation is equal to the termination threshold.

Images