A sample point evaluation process parallelization method for complex optimization design problems

Abstract

The application discloses a sample point evaluation process parallelization method for a complex optimization design problem and belongs to the technical field of aircraft aerodynamic configuration design. In order to improve the aircraft multidisciplinary performance evaluation efficiency and the multidisciplinary optimization design efficiency, the application comprises the following steps: drawing a multidisciplinary sample point evaluation flowchart, wherein the flowchart comprises the relationships among various disciplines; representing a single sample point evaluation process as a sub-process set; determining the levels of the sub-process set from left to right, setting the level of the leftmost sub-process as the first level, and marking the level of each sub-process in the order from left to right based on the increasing mode of the branch level; creating a layered task pool of each level, decomposing all sample evaluation processes into multiple irrelevant parallel task pools classified according to the levels; and completing the serial sample point parallelization evaluation processes of each layered task pool in the order of the levels one by one, and calculating the single sample point evaluation efficiency. The application can improve the aircraft multidisciplinary performance evaluation efficiency and the multidisciplinary optimization design efficiency.

Description

Technical Field

[0001] This invention belongs to the field of aircraft aerodynamic shape design technology, specifically relating to a parallel method for the sample point evaluation process for complex optimization design problems. Background Technology

[0002] Modern aircraft aerodynamic design primarily utilizes computational fluid dynamics and numerical optimization techniques, employing computers to automatically find the optimal aerodynamic shape that meets performance requirements. With the advancement of aircraft design technology, aerodynamic design has evolved from solely considering aerodynamic performance to addressing complex system design problems that comprehensively consider constraints and performance from disciplines such as structure, control, and flight mechanics, with aerodynamics at the core.

[0003] In multidisciplinary optimization design, the optimization algorithm iteratively generates various aerodynamic shapes for the aircraft and performs numerical analysis on the performance of each discipline to obtain an optimization path, thereby finding the optimal aerodynamic shape. In multidisciplinary optimization design, the process of generating the aerodynamic shape of the aircraft and evaluating its performance in each discipline is a single-sample-point evaluation process.

[0004] When dealing with complex optimization design problems, the sample point evaluation process involves performance evaluation across multiple disciplines. Data reuse exists between disciplines, and the logical sequence of the evaluation process is complex, making it a more complex sample point evaluation process compared to single-discipline optimization design. Current optimization design processes often employ a serial approach for sample point evaluation, utilizing one or more CPU cores to sequentially evaluate the performance of all disciplines for a single sample point. While this method reduces system complexity, it fails to consider interdisciplinary relationships. If there is no interaction between disciplines, it is still necessary to wait for other disciplines to be evaluated before proceeding with the evaluation. This method wastes computing resources, reduces the efficiency of multi-disciplinary performance evaluation for aircraft, and further reduces the overall efficiency of aerodynamic shape optimization design for the aircraft. Summary of the Invention

[0005] The problem this invention aims to solve is to improve the efficiency of multidisciplinary performance evaluation and multidisciplinary optimization design of aircraft, and proposes a parallel method for the sample point evaluation process for complex optimization design problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A parallelization method for the sample point evaluation process of complex optimization design problems includes the following steps:

[0008] S1. Draw a flowchart for the multidisciplinary sample point evaluation, including the relationships between the disciplines;

[0009] S2. Based on the multidisciplinary sample point evaluation flowchart obtained in step S1, the single sample point evaluation process is represented as a set of subprocesses;

[0010] S3. For the set of sub-processes obtained in step S2, determine the hierarchy from left to right, set the hierarchy of the leftmost sub-process as the first level, and label the hierarchy of each sub-process from left to right based on the increasing branch hierarchy.

[0011] S4. Create a hierarchical task pool for each level, and decompose the evaluation process of all samples into multiple unrelated and parallelizable task pools classified by level;

[0012] S5. Perform the parallel evaluation process of serial sample points for each hierarchical task pool in hierarchical order, and calculate the evaluation efficiency of a single sample point.

[0013] Furthermore, the multidisciplinary sample point evaluation in step S1 includes one or a combination of aerodynamic performance evaluation, structural performance evaluation, control performance evaluation, and flight mechanical performance evaluation.

[0014] Furthermore, step S2 represents the evaluation process of a single sample point as a set of sub-processes, expressed as:

[0015]

[0016] Where P is the set of subprocesses. Let i be the i-th subprocess.

[0017] Furthermore, the specific implementation method of step S3 is to hierarchically label the sub-process set P, setting the sub-process to contain three states: no prerequisite dependencies, prerequisite dependencies exist and the number of prerequisite dependencies is 1, and prerequisite dependencies exist and the number of prerequisite dependencies is greater than 1. The expression is:

[0018]

[0019] in, for The hierarchy express The set of all preceding subprocesses, express The number of elements in the set of all preceding subprocesses To find the maximum value.

[0020] Furthermore, the method for creating the hierarchical task pool for each level in step S4 is to set a set of hierarchical identifiers. Sample set number of floors Hierarchical task pool The expression for the hierarchical process is:

[0021]

[0022] in, For the first Layered task pool, For the first One sample, This is the current floor number.

[0023] Furthermore, in step S5, the evaluation efficiency for a single sample point is calculated. The calculation formula is:

[0024]

[0025] in, Indicates the serial evaluation time. Indicates the time for parallel evaluation. express Execution time, For the first Layered task pool.

[0026] The beneficial effects of this invention are:

[0027] This invention discloses a parallelization method for sample point evaluation in complex optimization design problems. By considering the multidisciplinary evaluation relationships within the samples, the sequential sample point evaluation process in complex optimization design is parallelized, thereby improving the efficiency of multidisciplinary performance evaluation and multidisciplinary optimization design for aircraft. This invention can be applied to the multidisciplinary optimization design process of aircraft to improve the efficiency of multidisciplinary performance evaluation and multidisciplinary optimization design. Attached Figure Description

[0028] Figure 1 This is a flowchart of a parallelization method for the sample point evaluation process for complex optimization design problems, as described in this invention.

[0029] Figure 2 This is a schematic diagram illustrating the multidisciplinary relationships and evaluation process in the multidisciplinary optimization design of this invention;

[0030] Figure 3 This is a schematic diagram illustrating the hierarchical identification of each sub-process in the evaluation process of this invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.

[0032] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.

[0033] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 -Appendix Figure 3 Detailed explanation is as follows:

[0034] Example 1:

[0035] A parallelization method for the sample point evaluation process of complex optimization design problems includes the following steps:

[0036] S1. Draw a flowchart for the multidisciplinary sample point evaluation, including the relationships between the disciplines;

[0037] Furthermore, the multidisciplinary sample point evaluation in step S1 includes one or a combination of aerodynamic performance evaluation, structural performance evaluation, control performance evaluation, and flight mechanical performance evaluation.

[0038] S2. Based on the multidisciplinary sample point evaluation flowchart obtained in step S1, the single sample point evaluation process is represented as a set of subprocesses;

[0039] Furthermore, step S2 represents the evaluation process of a single sample point as a set of sub-processes, expressed as:

[0040]

[0041] Where P is the set of subprocesses. This is the i-th subprocess;

[0042] S3. For the set of sub-processes obtained in step S2, determine the hierarchy from left to right, set the hierarchy of the leftmost sub-process as the first level, and label the hierarchy of each sub-process from left to right based on the increasing branch hierarchy.

[0043] Furthermore, the specific implementation method of step S3 is to hierarchically label the sub-process set P, setting the sub-process to contain three states: no prerequisite dependencies, prerequisite dependencies exist and the number of prerequisite dependencies is 1, and prerequisite dependencies exist and the number of prerequisite dependencies is greater than 1. The expression is:

[0044]

[0045] in, for The hierarchy express The set of all preceding subprocesses, express The number of elements in the set of all preceding subprocesses To find the maximum value.

[0046] S4. Create a hierarchical task pool for each level, and decompose the evaluation process of all samples into multiple unrelated and parallelizable task pools classified by level;

[0047] Furthermore, the method for creating the hierarchical task pool for each level in step S4 is to set a set of hierarchical identifiers. Sample set number of floors Hierarchical task pool The expression for the hierarchical process is:

[0048]

[0049] in, For the first Layered task pool, For the first One sample, This is the current floor number.

[0050] S5. Perform the parallel evaluation process of serial sample points for each hierarchical task pool in hierarchical order, and calculate the evaluation efficiency of a single sample point.

[0051] Furthermore, in step S5, the evaluation efficiency for a single sample point is calculated. The calculation formula is:

[0052]

[0053] in, Indicates the serial evaluation time. Indicates the time for parallel evaluation. express Execution time, For the first Layered task pool.

[0054] This implementation method is applied in practice as follows: Figure 2 As shown, first, the level of the leftmost terminal procedure is set to Lev_1, as follows: Figure 2 As shown in sub-process ①;

[0055] Assign values ​​to each sub-procedure in left-to-right order. When encountering branches, the branch level increases. Figure 2 From ① to ② and ③, ① is at level Lev_1, and ② and ③ are at level Lev_2;

[0056] In the case of no branches, they are at the same level, such as Figure 2 In steps ④ and ⑤, these two subprocesses should be combined during the evaluation; it should be noted that unconnected subprocesses at the same level should not be combined.

[0057] For a subprocess where multiple branches converge, the level of that subprocess is determined by adding 1 to the maximum value of all connected levels, such as... Figure 2 As shown in component ⑧, the level of component ⑧ is the maximum value of ⑤⑥⑦ plus 1, which is Lev_5.

[0058] Then, construct task pools at different levels, such as Figure 3 As shown, there are five task pools, from level 1 to level 5. Tasks within the same level are independent and can be completed in parallel. All tasks in each task pool are completed sequentially and in parallel, enabling parallel multidisciplinary performance evaluation across multiple sample evaluation processes and between disciplines unrelated to a single sample evaluation process. The process of constructing task pools at different levels involves aggregating different sample points to be evaluated processes with the same level identifier to form a set of processes to be evaluated at that level. Each process in the set has two identifiers: a level identifier and a sample sequence number identifier.

[0059] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0060] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for parallelizing sample point evaluation process for complex optimization design problems, characterized in that, Comprising the following steps: S1. Draw a multi-disciplinary sample point evaluation flow chart, which includes the relationship between disciplines; The multi-disciplinary sample point evaluation in step S1 includes one or a combination of aerodynamic performance evaluation, structural performance evaluation, control performance evaluation, and flight mechanics performance evaluation; S2. Based on the multi-disciplinary sample point evaluation flow chart obtained in step S1, represent the single sample point evaluation process as a set of sub-processes; Step S2 represents the single sample point evaluation process as a set of sub-processes, and the expression is: P = {p1, p2,..., pn} is a set of points in the plane. i ..., pn} is a set of points in the plane. n} wherein P is a set of sub-processes, p i is the i-th sub-process; S3. For the set of sub-processes obtained in step S2, determine the hierarchy from left to right, set the hierarchy of the leftmost sub-process to the first layer, and label the hierarchy of each sub-process in order from left to right based on the increasing manner of branch hierarchy; The specific implementation method of step S3 is to label the hierarchy of the set of sub-processes P, and set three states for the sub-processes, namely no pre-dependence, pre-dependence with a number of pre-dependences being 1, and pre-dependence with a number of pre-dependences being greater than 1, and the expression is: wherein Lev(p i ) is the level of p i , Pre(p i ) represents the set of all preceding sub-processes of p i , and represents the number of elements in the set of all preceding sub-processes of p i , and max is the maximum value. S4. Create a hierarchical task pool for each level, and decompose all sample evaluation processes into multiple independent parallel task pools classified by level; S5. Complete the serial sample point parallelization evaluation process of each hierarchical task pool in order of level, and calculate the single sample point evaluation efficiency.

2. The sample point evaluation process parallelization method for a complex optimization design problem according to claim 1, wherein, The method for creating the hierarchical task pool of each level in step S4 is to set the level identification set {Lev(p i )}, the sample set S, the number of layers K, and the hierarchical task pool {Pool1, Pool2, …, Pool K}. The expression of the hierarchical process is: where Pool k is the k-th layer task pool, s j is the j-th sample, and k is the current layer number.

3. The method of claim 2, wherein, The calculation formula of the single sample point evaluation efficiency η in step S5 is: where T serial represents the serial evaluation time, T parallel represents the parallel evaluation time, t i represents the execution time of p i , Pool k is the kth layer task pool.

Images