Furniture layout method, device, electronic device and storage medium

Abstract

A furniture layout method, equipment, storage medium and device, the method comprises the steps of: 1) parsing the layout space and furniture information input by the user to construct an initial layout plan; 2) dynamically adjusting the search window by using an adaptive window adjustment module according to the characteristics of the layout problem, and using the window to perform global or local optimization search based on the current layout plan to generate a set of new layout plans; 3) adaptively fine-tuning and correcting the newly generated layout plan by using a phase correction module, and further optimizing the layout plan by using a conflict detection and resolution mechanism, and adding the optimized layout plan to a candidate set; 4) repeatedly executing steps 2) and 3) until the iteration termination condition is met; 5) comprehensively evaluating the candidate layout set by using a Pareto optimization strategy, and selecting the optimal layout plan as output.

Description

Technical Field

[0001] The present invention relates to the technical field of furniture layout, and more specifically, to a furniture layout method, device, electronic equipment and storage medium. Background Art

[0002] In recent years, with the rapid development of computer-aided design (CAD) technology and the continuous advancement of artificial intelligence (AI) algorithms, the automation and intelligence level of furniture layout design has been significantly improved. At present, a variety of methods based on different optimization strategies have been proposed and applied, such as: layout optimization based on genetic algorithm (GA), layout optimization based on particle swarm optimization (PSO), layout optimization based on ant colony algorithm (ACO), etc. These methods have improved the efficiency and quality of furniture layout design to a certain extent and reduced the workload of designers.

[0003] However, existing methods still have the following shortcomings when dealing with complex layout problems:

[0004] (1) The local search capability of the algorithm is limited. Most existing methods use a single optimization algorithm, such as GA and PSO, which perform well in global search capabilities but are relatively weak in local search capabilities. Taking GA as an example, since its search mainly relies on random operations such as gene crossover and mutation, it lacks an effective guidance strategy in the local search stage and is prone to fall into the local optimum and is difficult to further optimize. This makes it difficult for the layout solution to achieve the optimal local details, affecting the layout quality.

[0005] (2) Imperfect multi-objective trade-off and optimization mechanism. Furniture layout design usually involves multiple optimization objectives, such as space utilization, aesthetics, comfort, etc. There are certain conflicts and trade-offs between these objectives. Most existing methods use weighted summation to transform multiple objectives into a single objective for optimization, or use subjective weighting methods such as the analytic hierarchy process (AHP) to determine the objective weights. This processing method is difficult to fully describe the complex trade-off relationship between objectives, and lacks an objective and efficient multi-objective optimization mechanism, which affects the comprehensive performance of the layout scheme.

[0006] (3) Insufficient integration of domain knowledge. Furniture layout design needs to consider many professional specifications and design criteria, such as the safety of furniture placement, convenience of passage, rationality of lighting, etc. Existing methods do not integrate these domain knowledge sufficiently during the optimization process, which is mainly manifested in two aspects: first, the representation and modeling of domain knowledge is not systematic and standardized enough, making it difficult to accurately describe design rules; second, the constraint processing of domain knowledge in the optimization algorithm is not flexible enough, which easily leads to a layout plan that is too simple or infeasible. This limits the professionalism and practicality of the layout plan.

[0007] (4) Weak user interaction and adaptive capabilities. Furniture layout design serves the personalized needs of users and needs to fully consider user preferences and feedback. However, existing methods do not provide sufficient support for user interaction during the optimization process and lack effective user preference guidance and feedback update mechanisms, which makes it difficult for layout solutions to meet the diverse needs of users. At the same time, existing algorithms generally lack adaptive tuning capabilities and are difficult to dynamically adjust strategies based on the characteristics of the problem and feedback information from the optimization process, which affects the generalization ability of the algorithm and the actual application effect.

[0008] To address the above problems, a new method is urgently needed that can improve the local search capability of the algorithm while introducing a more scientific and efficient multi-objective optimization mechanism, achieve a deep integration of domain knowledge and optimization algorithms, and support user interaction and algorithm adaptation, in order to obtain a more professional, practical and personalized furniture layout design solution. Summary of the invention

[0009] In view of the deficiencies of the prior art, the present invention proposes a furniture layout design method based on adaptive window search and Pareto optimization. By introducing a variety of intelligent algorithms at different stages of optimization, a collaborative optimization mechanism with simulated annealing as the main framework and Pareto optimization as the auxiliary is constructed, which effectively balances the global search and local search capabilities of the algorithm. At the same time, the layout phase decomposition and constraint graph modeling strategies are adopted to realize the systematic representation and flexible integration of domain knowledge, and the dynamic integration of user interaction and algorithm adaptation is realized through Pareto frontier presentation and interactive feedback guidance. This method has better solved the deficiencies of the above-mentioned prior art and has important application value and promotion potential.

[0010] The present invention provides a furniture layout method, comprising the following steps:

[0011] 1) Parse the layout space and furniture information input by the user to construct an initial layout plan; 2) According to the characteristics of the layout problem, use the adaptive window adjustment module to dynamically adjust the search window, and use the window to perform global or local optimization search based on the current layout plan to generate a set of new layout plans; 3) Use the phase correction module to adaptively fine-tune and correct the newly generated layout plan, and further optimize the layout plan through the conflict detection and resolution mechanism, and add the optimized layout plan to the candidate set; 4) Repeat steps 2) and 3) until the iteration termination condition is met; 5) Use the Pareto optimization strategy to comprehensively evaluate the candidate layout set and select the optimal layout plan as the output.

[0012] Specifically, the adaptive window adjustment module in step 2) adopts an adaptive window search algorithm SAWS based on simulated annealing, including the following steps:

[0013] 2.1) Initialize the size factor α and shape factor β of the search window, as well as the annealing temperature T, where α ∈ (0, 1), representing the size ratio of the search window to the entire layout space, with the initial value α 0 = 0.5; $$, representing the degree of irregularity of the search window shape, with the initial value β 0 = 0; T is the temperature parameter of the simulated annealing algorithm, with the initial value T 0 = 100;

[0014] 2.2) Based on the current layout scheme, randomly select a group of furniture using the search window and perform local perturbations on their positions and orientations to obtain a new local layout L':

[0015] Position perturbation: Inside the search window, for each piece of furniture f i ∈ F', its position coordinates (x i , y i ) are randomly offset by a step size δ, i.e., x' i = x i + rand(-δ, δ), y' i = y i + rand(-δ, δ), where rand(-δ, δ) represents random sampling within the range [-δ, δ];

[0016] Orientation perturbation: For each piece of furniture f i ∈ F', its orientation angle θ i is randomly deflected by a step size η, i.e., θ' i = θ i + rand(-η, η);

[0017] 2.3) Use the layout evaluation function E(·) to score the new layout L', and decide whether to accept L' as the new current layout according to the Metropolis criterion; if E(L') < E(L), then accept L' as the new current layout, otherwise, accept L' as the new current layout with a probability P = exp[-(E(L') - E(L)) / T], where E(L) and E(L') are the evaluation function values of the current layout and the new layout respectively;

[0018] 2.4) According to the current iteration number t and the maximum iteration number T max , update the size factor α and shape factor β of the search window:

[0019] α = α 0 × (1 - tanh(k 1 × t / T max ))

[0020] β = β 0 + (1 - β0 )×sin(k 2 ×π×t / T max )

[0021] Straight K 1 is the resizing rate factor, k 2 Adjust the rate factor for shape;

[0022] 2.5) According to the current number of iterations t and the maximum number of iterations T max , update the annealing temperature T

[0023] T=T 0 ×exp(-k 3 ×t / T max );

[0024] where k 3 is the cooling rate factor;

[0025] 2.6) Repeat steps 2.2) to 2.5) until the iteration termination condition is met.

[0026] Specifically, the phase correction module in step 3) adopts an adaptive correction algorithm LPAC based on layout phase analysis, including the following steps:

[0027] 3.1) Perform phase decomposition on the furniture layout in the candidate layout scheme L and extract the following key phase components:

[0028] Furniture spacing D = {d ij |i, j∈F, i≠j}, where d ij Represents the distance between furniture i and j, calculated as: d ij =[(x i -x j ) 2 +(y i -y j ) 2 ] 0.5 ;

[0029] Alignment A = {a ij |i,j∈F,i≠j}, where a ij It indicates the alignment degree of furniture i and j in the x direction or y direction, and the calculation formula is: ij =min(|x i -x j |,|y i -y j |);

[0030] Orientation consistency O = {o i |i∈F}, where o i represents the orientation angle θ of furniture ii The angle between the main direction of the room is calculated as: i =min(|θ i -θ 0 |, 180°-|θ i -θ 0 |), θ 0 is the main direction angle of the room;

[0031] 3.2) Adaptively correct each phase component to obtain the corrected layout L′:

[0032] Spacing correction: For each pair of furniture (i, j), calculate the corrected spacing d′ ij =d ij +λ 1 ·sign(d ij -d 0 )·log(1+|d ij -d 0 |), where d 0 is the spacing reference value λ 1 is the spacing correction strength factor, sign(·) is the sign function;

[0033] Alignment correction: For each pair of furniture (i, j), calculate its corrected alignment a′ ij =a ij +λ 2 ·sin(π·a ij / a 0 )·exp(-a ij / a 0 ), where a 0 is the reference value of alignment, λ 2 Correction intensity factor for alignment;

[0034] Orientation correction: For each piece of furniture i, calculate its corrected orientation angle θ′ i =θ i +λ 3 ·tanh(o i / o 0 )·cos(2·o i ), where o 0 is the reference value for orientation consistency, λ 3 is the orientation correction intensity factor;

[0035] 3.3) Reconstruct the layout L′ according to the corrected phase component and add it to the candidate set C as a new candidate layout.

[0036] Specifically, after the phase correction, the conflict detection and resolution of the layout scheme using a heuristic search algorithm based on a constraint graph includes the following steps:

[0037] 4.1) Construct the layout constraint graph G = (V, E), where V is the set of furniture nodes and E is the set of constraint edges

[0038] V={v i |i∈F}, where v i Indicates the node corresponding to furniture i

[0039] E={e ij |i, j∈F, i≠j, constraint (i, j) = True}, where e ij Represents node v i and v j The constraint edge between them, constraint (i, j) is a Boolean function indicating whether there is a layout constraint between furniture i and j;

[0040] 4.2) Define the layout constraint rule set R. Common constraint rules include:

[0041] Furniture does not overlap: Where Inter(i,j) is a Boolean function indicating whether furniture i and j have an intersection;

[0042] Reasonable distance from the wall: Where d(i, W) represents the distance from furniture i to the nearest wall, d min and d max is the lower and upper distance limit;

[0043] Accessibility: If there is a path from i to j, then Width(i, j) ≥ w minν Where Width(i, j) represents the minimum width of the path, w min is the lower limit of width;

[0044] 4.3) Use the depth-first search algorithm to traverse the constraint graph G and mark the conflicting nodes and conflicting edges. Conflicting nodes: Make v i Violation of any rule r∈R; conflicting edge: Make e ij Violation of any rule r∈R;

[0045] 4.4) For each conflicting node v i , perform heuristic adjustment according to the rule set R to obtain the adjusted furniture layout L′; if v i If the furniture does not overlap, a new position is randomly sampled in its neighborhood until it no longer overlaps. i If the "reasonable distance from the wall" rule is violated, move v in a direction perpendicular to the nearest wall. i , until the distance limit is met; if v iIf the "convenient passage" rule is violated, choose a suitable passage and adjust the position of related furniture so that the passage meets the width requirements;

[0046] 4.5) For each conflicting edge e ij , perform heuristic adjustment according to the rule set R and obtain the adjusted furniture layout L′: if e ij If the "furniture does not overlap" rule is violated, adjust v separately. i and v j position until there is no overlap; if e ij If the "reasonable wall distance" rule is violated, adjust v i and v j positions so that they all satisfy the distance restriction; if e ij If the "traffic convenience" rule is violated, adjust v i and v j The location and layout of the passages make sure that the passages meet the width requirements;

[0047] 4.6) Add the adjusted layout L′ as a new candidate layout to the candidate set C.

[0048] Specifically, the iteration termination condition in step 4) includes reaching the maximum number of iterations T max ; The optimal layout score improvement of K consecutive iterations is less than the threshold ∈; The current temperature T is lower than the set minimum temperature T min .

[0049] Specifically, the step 5) adopts a multi-objective layout evaluation and selection algorithm PMOES based on Pareto optimization, including the following steps:

[0050] 6.1) Define a set of layout evaluation indicators U = {u 1 ,u 2 , ..., u n}, for each layout solution L in the candidate set C i , calculate its score on each indicator (u i1 ,u i2 , ..., u in ):

[0051] Space utilization: u i1 =U si / S, where U si For layout L i The total area occupied by the furniture in the room, S is the total area of ​​the room;

[0052] Lighting index: u i2 =W i / W, where W i For layout L iThe total area of ​​all windows W is the total window area of ​​the room;

[0053] Aesthetics: u i3 =(Sym i +Bal i +Div i ) / 3, where Sym i For layout L i Symmetry score, Bal i For layout L i Balance score of Div i For layout L i The richness score of , the score range is [0, 1];

[0054] Comfort: where w j is the weight of the j-th comfort factor, g j is the actual value of the jth comfort factor, g j0 is the reference value of the jth comfort factor, and m is the total number of comfort factors;

[0055] Personalization: i5 Given by users through questionnaires or preference feedback, reflecting the layout L i The degree of matching with the user's personalized needs, with a value range of [0, 1];

[0056] 6.2) Based on the Pareto dominance relationship, the candidate set C is sorted in a non-inferior manner to obtain a set of Pareto frontiers:

[0057] Layout L i Pareto Dominant Layout L j′ If and only if and Layout L i With layout L j do not dominate each other if and only if and u il jl ; Pareto frontier

[0058] 6.3) Apply the weighted sum method on the Pareto frontier F, combining the user preference weight vector w = (w 1 , w 2 , ..., w n ), calculate each layout L i Overall ratings i :

[0059]

[0060] where w​k is the weight of the kth indicator,

[0061] 6.4) Select the layout with the highest comprehensive score as the optimal recommended layout L * :

[0062]

[0063] 6.5) The optimal recommended layout L * It is returned together with other candidate layouts in the Pareto frontier F for user reference and selection;

[0064] 6.6) Record user feedback on the recommended layout, dynamically update the preference weight vector w, and achieve continuous optimization of layout recommendations.

[0065] Specifically, the layout space includes a room plan and a furniture list. The plan uses Canny edge detection and Hough transform to extract room boundaries and door and window positions. The furniture is represented parameterized, and the parameters include: center coordinates (x, y), length l, width w and orientation θ.

[0066] A furniture layout device, comprising: a layout parsing module, an adaptive window search module, a layout phase correction module, a layout evaluation and selection module, and a human-computer interaction module;

[0067] The layout parsing module is used to parse the layout space and furniture information input by the user and construct an initial layout plan;

[0068] The adaptive window search module is used to dynamically adjust the search window according to the characteristics of the layout problem, and use the window to perform global or local optimization search based on the current layout solution to generate a set of new layout solutions;

[0069] The layout phase correction module is used to adaptively fine-tune and correct the newly generated layout scheme, and further optimize it through the conflict detection and resolution mechanism, and add the optimized layout scheme to the candidate set;

[0070] The layout evaluation and selection module is used to comprehensively evaluate the candidate layout set using the Pareto optimization strategy and select the optimal layout solution as output;

[0071] The human-computer interaction module is used to receive layout space and furniture information input by the user, and present the optimized layout plan to the user.

[0072] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the computer program.

[0073] A storage medium stores a computer program, wherein the computer program implements the steps of any of the above methods when executed by a processor.

[0074] The present invention has the following beneficial effects:

[0075] 1. The present invention cleverly integrates multiple intelligent optimization algorithms such as simulated annealing (SA), Pareto optimization (PO), and constraint satisfaction (CS) into different stages of layout optimization, forming a collaborative optimization algorithm framework. Among them, the adaptive window search module uses the SA algorithm to realize the global exploration and local refinement of the layout plan, and balances the breadth and depth of the search by adaptively adjusting the search window. The layout phase correction module uses the PO algorithm to realize the multi-objective trade-off and optimization of the layout plan, and improves the comprehensive performance of the layout plan through the construction of the Pareto Front. The constraint graph conflict resolution module uses the CS algorithm to realize automatic error correction and optimization of the layout plan, and finds a feasible solution that satisfies complex constraints through heuristic search. These three modules have clear division of labor and complementary advantages. They play a synergistic role in different stages of layout optimization and avoid the limitations of a single algorithm in complex layout problems.

[0076] 2. The furniture layout problem involves complex factors in multiple dimensions, such as layout space, furniture parameters, design specifications, and user needs, which are difficult to model and solve directly. The present invention adopts the idea of ​​layout phase decomposition, abstracting the layout problem into multiple key phase components, such as furniture spacing, alignment, orientation consistency, etc., and performs targeted optimization on the basis of problem decoupling, thereby simplifying the complexity of the algorithm. At the same time, through the definition of layout evaluation indicators, the qualitative layout requirements are transformed into quantitative optimization goals, realizing the multi-criteria characterization and evaluation of the layout problem. In addition, the introduction of user preference weights in Pareto optimization can flexibly balance the performance of various layout solutions to meet the personalized needs of users. These strategies characterize the inherent structure and key influencing factors of the layout problem from multiple perspectives, and improve the pertinence and interpretability of layout optimization.

[0077] 3. Furniture layout design needs to follow certain professional standards and design common sense, such as: furniture does not overlap, reasonable wall distance, convenient access, etc. The present invention uses constraint graphs to explicitly model these domain knowledge, converts qualitative design rules into quantitative constraints, and realizes the computable representation of domain knowledge. On this basis, through flexible constraint processing techniques, such as penalty functions, heuristic searches, etc., a layout scheme with greater flexibility under the premise of meeting the constraints is found, avoiding the problem of layout uniformity caused by overly rigid constraints. In addition, soft evaluation factors such as comfort and aesthetics are included in the layout evaluation indicators to guide the layout scheme to take into account the user experience and aesthetic quality while meeting the hard specifications. These strategies fully reflect the professional characteristics of furniture layout design in layout optimization and improve the professionalism and practicality of the layout scheme.

[0078] 4. The present invention introduces multiple adaptive adjustment mechanisms in the process of layout optimization, such as dynamic changes in the search window, adaptive fine-tuning of phase correction, adaptive attenuation of annealing parameters, etc., so that the algorithm can dynamically adjust the strategy according to the feedback information of the optimization process, thereby improving the algorithm's adaptive ability to different layout problems. At the same time, through the presentation of Pareto Front and the dynamic update of user preferences, the dynamic integration of layout optimization and user interaction is achieved. On the one hand, the algorithm adjusts the optimization strategy according to user feedback to make the layout plan more in line with user needs; on the other hand, the user understands the design space and performance trade-offs of the layout plan through the visualized Pareto Front, thus achieving human-computer collaborative optimization. This mechanism combining adaptive optimization with user interaction improves the practicality and user experience of layout optimization.

[0079] The method of the present invention is highly innovative in the existing field of furniture layout optimization, breaking through the limitations of traditional optimization algorithms on complex layout problems, and realizing the organic integration and synergistic efficiency of multiple intelligent optimization technologies. From the system level, the division of labor and cooperation of multiple modules and the unification of interfaces improve the solving ability and generalization ability of the layout optimization system. From the algorithm level, the efficiency, robustness and interpretability of the layout optimization algorithm are improved through strategies such as problem decoupling, domain modeling, and adaptive tuning. In addition, the present invention also takes into account factors such as user interaction and professional specifications, making layout optimization closer to actual application scenarios. These technical features give the present invention broad application prospects in the automation and intelligence of furniture layout design. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 is a flow chart of the furniture layout method of the present invention;

[0081] Figure 2 It is a schematic diagram of the furniture layout device of the present invention. DETAILED DESCRIPTION

[0082] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the furniture layout method proposed by the present invention, its specific implementation method, structure, features and effects are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0083] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0084] Example 1

[0085] See also Figure 1 , a furniture layout method, comprising the following steps:

[0086] Step 1. Layout space and furniture information analysis:

[0087] Parse the layout space and furniture information input by the user to build an initial layout plan. The layout space includes the room floor plan and the furniture list. The floor plan uses the Canny edge detection algorithm to identify the room outline, and the Hough transform is used to extract the door and window positions. The furniture is represented by parameters, including the type, size, placement requirements, etc. Each piece of furniture is represented by a parameter vector, such as (ID, x, y, l, w, θ) to represent the furniture number, center coordinates, length, width and orientation. Normalize the furniture parameters to the [0, 1] interval for subsequent processing. Based on the room size and furniture parameters, a set of initial layouts are generated using random sampling or predefined rules.

[0088] Step 2. Adaptive window adjustment and layout search:

[0089] According to the characteristics of the layout problem, the adaptive window adjustment module is used to dynamically adjust the search window, and the window is used to perform global or local optimization search based on the current layout solution to generate a set of new layout solutions. This module uses the adaptive window search algorithm SAWS based on simulated annealing, and balances global exploration and local utilization by dynamically adjusting the size and shape of the search window, thereby improving the efficiency and quality of layout optimization.

[0090] The adaptive window adjustment module in step 2) adopts an adaptive window search algorithm SAWS based on simulated annealing, including the following steps:

[0091] 2.1) Initialize the size factor α and shape factor β of the search window, as well as the annealing temperature T, where α ∈ (0, 1), representing the size ratio of the search window to the entire layout space, and the initial value of α 0 = 0.5; β represents the degree of irregularity of the search window shape, and the initial value of β 0 = 0; T is the temperature parameter of the simulated annealing algorithm, and the initial value of T 0 = 100;

[0092] 2.2) Based on the current layout scheme, randomly select a group of furniture using the search window and perform local perturbations on their positions and orientations to obtain a new local layout L':

[0093] Position perturbation: Inside the search window, for each piece of furniture f i ∈ F', its position coordinates (x i , y i ) are randomly offset by a step size δ, that is, x' i = x i + rand(-δ, δ), y' i = y i + rand(-δ, δ), where rand(-δ, δ) represents random sampling within the range [-δ, δ];

[0094] Orientation perturbation: For each piece of furniture f i ∈ F', its orientation angle θ i is randomly deflected by a step size η, that is, θ' i = θ i + rand(-η, η);

[0095] 2.3) Use the layout evaluation function E(·) to score the new layout L', and decide whether to accept L' as the new current layout according to the Metropolis criterion; if E(L') < E(L), then accept L' as the new current layout, otherwise, accept L' as the new current layout with a probability P = exp[-(E(L') - E(L)) / T], where E(L) and E(L') are the evaluation function values of the current layout and the new layout respectively;

[0096] 2.4) According to the current iteration number t and the maximum iteration number T max , update the size factor α and shape factor β of the search window:

[0097] α = α 0 × (1 - tanh(k 1 × t / T max ))

[0098] β = β 0 + (1 - β0 )×sin(k 2 ×π×t / T max )

[0099] Straight K 1 is the resizing rate factor, k 2 Adjust the rate factor for shape;

[0100] 2.5) According to the current number of iterations t and the maximum number of iterations T max , update the annealing temperature T

[0101] T=T 0 ×exp(-k 3 ×t / T max );

[0102] where k 3 is the cooling rate factor;

[0103] 2.6) Repeat steps 2.2) to 2.5) until the iteration termination condition is met.

[0104] In the specific implementation, the initial size factor α of the search window is 0 It can be around 0.5, the initial shape factor β 0 Can be 0, annealing initial temperature T 0 It can be around 100, and the maximum number of iterations is T max It can be set appropriately according to the complexity of the problem, usually 1000 to 5000. Parameter adjustment rate factor k 1 , k 2 , k 3 The value of needs to be selected empirically based on the specific problem, and is usually in the range of 0.01 to 0.1.

[0105] Step 3. Layout phase correction and conflict resolution:

[0106] The newly generated layout plan is adaptively fine-tuned and corrected using the phase correction module, and the layout plan is further optimized through the conflict detection and resolution mechanism, and the optimized layout plan is added to the candidate set. The phase correction module uses the adaptive correction algorithm LPAC based on layout phase analysis to decompose and correct the position, direction and other elements in the furniture layout, identify and resolve layout conflicts, and improve the rationality and aesthetics of the layout plan.

[0107] The phase correction module in step 3 adopts the adaptive correction algorithm LPAC based on layout phase analysis, including the following steps:

[0108] 3.1) Perform phase decomposition on the furniture layout in the candidate layout scheme L and extract the following key phase components:

[0109] Furniture spacing D = {d ij |i, j∈F, i≠j}, where d ij Represents the distance between furniture i and j, calculated as: d ij =[(x i -x j ) 2 +(y i -y j ) 2 ] 0.5 ;

[0110] Alignment A = {a ij |i,j∈F,i≠j}, where a ij It indicates the alignment degree of furniture i and j in the x direction or y direction, and the calculation formula is: ij =min(|x i -x j |,|y i -y j |);

[0111] Orientation consistency O = {o i |i∈F}, where o i represents the orientation angle θ of furniture i i The angle between the main direction of the room is calculated as: i =min(|θ i -θ 0 |, 180°-|θ i -θ 0 |), θ 0 is the main direction angle of the room;

[0112] 3.2) Adaptively correct each phase component to obtain the corrected layout L′:

[0113] Spacing correction: For each pair of furniture (i, j), calculate the corrected spacing d′ ij =d ij +λ 1 ·sign(d ij -d 0 )·log(1+|d ij -d 0 |), where d 0 is the spacing reference value λ 1 is the spacing correction strength factor, sign(·) is the sign function;

[0114] Alignment correction: For each pair of furniture (i, j), calculate its corrected alignment a′ ij =a ij +λ 2 ·sin(π·a ij / a 0 )·exp(-a ij / a 0 ), where a 0 is the reference value of alignment, λ 2 Correction intensity factor for alignment;

[0115] Orientation correction: For each piece of furniture i, calculate its corrected orientation angle θ′ i =θ i +λ 3 ·tanh(o i / o 0 )·cos(2·o i ), where o 0 is the reference value for orientation consistency, λ 3 is the orientation correction intensity factor;

[0116] 3.3) Reconstruct the layout L′ according to the corrected phase component and add it to the candidate set C as a new candidate layout.

[0117] The algorithm flow is as follows:

[0118] (1) Perform phase decomposition on the furniture layout in the candidate layout scheme and extract key layout phase components, including furniture spacing D, alignment A, and orientation consistency O.

[0119] (2) Adaptively correct each phase component to obtain a corrected layout. Spacing correction is achieved through a logarithmic function, alignment correction is achieved through a sine function, and orientation correction is achieved through a hyperbolic tangent and cosine function. For specific correction formulas, see claim 3. Correction intensity factors λ1, λ2, and λ3 control the correction amplitude of each component, and are usually in the range of 0.1 to 1.

[0120] (3) Reconstruct the layout according to the corrected phase component and add it to the candidate set as a new candidate layout.

[0121] After the layout phase correction, it is necessary to use a heuristic search algorithm based on the constraint graph to detect and resolve conflicts in the layout scheme to further improve the feasibility of the layout. The following steps are included:

[0122] 4.1) Construct the layout constraint graph G = (V, E), where V is the set of furniture nodes and E is the set of constraint edges

[0123] V={v i |i∈F}, where v i Indicates the node corresponding to furniture i

[0124] E={e ij |i, j∈F, i≠j, constraint (i, j) = True}, where e ijRepresents node v i and v j The constraint edge between them, constraint (i, j) is a Boolean function indicating whether there is a layout constraint between furniture i and j;

[0125] 4.2) Define the layout constraint rule set R. Common constraint rules include:

[0126] Furniture does not overlap: Where Inter(i,j) is a Boolean function indicating whether furniture i and j have an intersection;

[0127] Reasonable distance from the wall: Where d(i, W) represents the distance from furniture i to the nearest wall, d min and d max is the lower and upper distance limit;

[0128] Accessibility: If there is a path from i to j, then Width(i, j) ≥ w minν Where Width(i, j) represents the minimum width of the path, w min is the lower limit of width;

[0129] 4.3) Use the depth-first search algorithm to traverse the constraint graph G and mark the conflicting nodes and conflicting edges. Conflicting nodes: Make v i Violation of any rule r∈R; conflicting edge: Make e ij Violation of any rule r∈R;

[0130] 4.4) For each conflicting node v i , perform heuristic adjustment according to the rule set R and obtain the adjusted furniture layout L′; if v i If the furniture does not overlap, a new position is randomly sampled in its neighborhood until it no longer overlaps. i If the "reasonable distance from the wall" rule is violated, move v in a direction perpendicular to the nearest wall. i , until the distance limit is met; if v i If the "convenient passage" rule is violated, choose a suitable passage and adjust the position of related furniture so that the passage meets the width requirements;

[0131] 4.5) For each conflicting edge e ij , perform heuristic adjustment according to the rule set R and obtain the adjusted furniture layout L′: if e ij If the "furniture does not overlap" rule is violated, adjust v separately. i and v j position until there is no overlap; if e ijIf the "reasonable wall distance" rule is violated, adjust v i and v j positions so that they all satisfy the distance restriction; if e ij If the "traffic convenience" rule is violated, adjust v i and v j The location and layout of the passages make sure that the passages meet the width requirements;

[0132] 4.6) Add the adjusted layout L′ as a new candidate layout to the candidate set C.

[0133] Specifically, the iteration termination condition in step 4) includes reaching the maximum number of iterations T max ; The optimal layout score improvement of K consecutive iterations is less than the threshold ∈; The current temperature T is lower than the set minimum temperature T min .

[0134] The algorithm flow is as follows:

[0135] (1) Construct a layout constraint graph G = (V, E), where V is the set of furniture nodes and E is the set of constraint edges.

[0136] (2) Define a set of layout constraint rules R. Common rules include: no furniture overlap, reasonable wall distance, convenient passage, etc.

[0137] (3) Use the depth-first search algorithm to traverse the constraint graph G and mark the conflicting nodes and conflicting edges.

[0138] (4) For each conflicting node, heuristic adjustments are made according to the rule set R, such as randomly sampling positions in the neighborhood to eliminate overlap, moving furniture in a direction perpendicular to the wall to meet distance restrictions, and adjusting the width of the passage to meet traffic requirements.

[0139] (5) For each conflicting edge, heuristic adjustments are made according to the rule set R to adjust the positions of related furniture and the layout of pathways to resolve the conflict.

[0140] (6) The adjusted layout is added to the candidate set as a new candidate layout.

[0141] In practical applications, layout constraint rules can be flexibly set according to the constraints and optimization goals of specific problems, such as defining overlap judgment thresholds, wall distance ranges, lower limits of passage widths, and other parameters to balance the feasibility and diversity of layout options.

[0142] Step 4. Iterative search and termination condition:

[0143] Repeat steps 2 and 3 to continuously update the candidate layout set and the current optimal layout until the iteration termination condition is met. The iteration termination condition in step 4) includes reaching the maximum number of iterations T max; The optimal layout score improvement of K consecutive iterations is less than the threshold ∈; The current temperature T is lower than the set minimum temperature T min .

[0144] During the iterative search process, the convergence and optimization effect of the algorithm can be monitored in real time by recording statistical information of the iterative process, such as the average layout score, the optimal layout score, the score improvement, etc. If necessary, the algorithm parameters can be dynamically adjusted, such as the annealing rate, search window parameters, etc., to further improve the efficiency and quality of layout optimization.

[0145] Step 5. Pareto layout evaluation and selection:

[0146] The Pareto optimization strategy is used to comprehensively evaluate the candidate layout set and select the optimal layout solution as the output. This step uses the multi-objective layout evaluation and selection algorithm PMOES based on Pareto optimization, comprehensively considers multiple evaluation indicators such as the practicality and aesthetics of the layout solution, and realizes the optimization and recommendation of the layout solution by constructing the Pareto Front.

[0147] Specifically, the step 5) adopts a multi-objective layout evaluation and selection algorithm PMOES based on Pareto optimization, including the following steps:

[0148] 6.1) Define a set of layout evaluation indicators U = {u 1 ,u 2 , ..., u n}, for each layout solution L in the candidate set C i , calculate its score on each indicator (u i1 ,u i2 , ..., u in ):

[0149] Space utilization: u i1 =U si / S, where U si For layout L i The total area occupied by the furniture in the room, S is the total area of ​​the room;

[0150] Lighting index: u i2 =W i / W, where W i For layout L i The total area of ​​all windows W is the total window area of ​​the room;

[0151] Aesthetics: u i3 =(Sym i +Bal i +Div i ) / 3, where Sym i For layout Li Symmetry score, Bal i For layout L i Balance score of Div i For layout L i The richness score of , the score range is [0, 1];

[0152] Comfort: where w j is the weight of the j-th comfort factor, g j is the actual value of the jth comfort factor, g j0 is the reference value of the jth comfort factor, and m is the total number of comfort factors;

[0153] Personalization: i5 Given by users through questionnaires or preference feedback, reflecting the layout L i The degree of matching with the user's personalized needs, with a value range of [0, 1];

[0154] 6.2) Based on the Pareto dominance relationship, the candidate set C is sorted in a non-inferior manner to obtain a set of Pareto frontiers:

[0155] Layout L i Pareto Dominant Layout L j′ If and only if and Layout L i With layout L j do not dominate each other if and only if Andu il jl ; Pareto frontier

[0156] 6.3) Apply the weighted sum method on the Pareto frontier F, combining the user preference weight vector w = (w 1 , w 2 , ..., w n ), calculate each layout L i Overall ratings i :

[0157]

[0158] where w k is the weight of the kth indicator,

[0159] 6.4) Select the layout with the highest comprehensive score as the optimal recommended layout L * :

[0160] ​

[0161] 6.5) The optimal recommended layout L * It is returned together with other candidate layouts in the Pareto frontier F for user reference and selection;

[0162] 6.6) Record user feedback on the recommended layout, dynamically update the preference weight vector w, and achieve continuous optimization of layout recommendations.

[0163] The algorithm flow is as follows:

[0164] (1) Define a set of layout evaluation indicators U = {u 1 ,u 2 , ..., u n}, for each layout scheme Li in the candidate layout set, calculate its score on various indicators. Layout indicators usually include: space utilization, lighting index, aesthetics, comfort, personalization, etc., which need to be selected and weighted according to actual application scenarios and user needs.

[0165] (2) Based on the Pareto dominance relationship, the candidate layout set is sorted in a non-inferior manner to obtain a set of Pareto frontiers. Non-inferior sorting continuously selects non-inferior solutions from the candidate set until no new non-inferior solutions can be found. The sorting result is the Pareto frontier.

[0166] (3) Apply the weighted summation method on the Pareto frontier, combine the user preference weight vector to calculate the comprehensive score of each layout scheme, and select the layout with the highest score as the optimal recommended layout. The user preference weight vector can be obtained through user portrait analysis, preference feedback, etc., and is continuously updated during the layout recommendation process to achieve personalized layout optimization.

[0167] (4) The optimal recommended layout is returned together with other candidate layouts in the Pareto frontier for user reference and selection. At the same time, the user's feedback on the recommended layout is recorded, and the preference weight is dynamically updated to achieve continuous optimization of layout recommendations.

[0168] In the specific implementation, the selection and weight setting of layout evaluation indicators are the key to Pareto optimization. Usually 3 to 5 representative indicators can be selected, and the weight setting needs to take into account the importance of each indicator and user preferences. The initial weight can be set based on experience or research, such as: space utilization 0.3, lighting index 0.2, aesthetics 0.2, comfort 0.2, personalization 0.1, and then dynamically adjusted according to user feedback. The generation of the Pareto frontier can adopt classic algorithms such as fast non-dominated sorting and congestion sorting, taking into account sorting efficiency and solution set diversity.

[0169] Through innovative algorithms such as adaptive window search, layout phase correction, and constraint graph conflict resolution, combined with Pareto multi-objective optimization and user preference guidance, a complete method is formed. This method can automatically generate practical, beautiful, and personalized optimal layout solutions while meeting design specifications and user needs, and provide a variety of candidate layouts for users to choose from, greatly improving the efficiency of furniture layout design and user experience.

[0170] The method of the present invention can be applied to various furniture layout design software and systems to provide users with intelligent layout optimization services. In addition, the core algorithm of the present invention has certain versatility and has broad application prospects in the fields of interior design, office space planning, cabin layout optimization, etc.

[0171] It should be noted that the algorithm parameter settings in the scheme of the present invention are only examples, and can be appropriately adjusted according to specific problems and data characteristics in actual application. In addition, the various steps and modules can be appropriately combined and split according to actual needs to optimize the modularity and scalability of the algorithm.

[0172] Example 2

[0173] Please refer to Figure 2 , a furniture layout device, comprising: a layout parsing module 1, an adaptive window searching module 2, a layout phase correction module 3, a layout evaluation and selection module 4 and a human-computer interaction module 5;

[0174] Layout parsing module 1, used to parse the layout space and furniture information input by the user and construct an initial layout plan;

[0175] Adaptive window search module 2, used to dynamically adjust the search window according to the characteristics of the layout problem, and use the window to perform global or local optimization search based on the current layout solution to generate a set of new layout solutions;

[0176] The layout phase correction module 3 is used to adaptively fine-tune and correct the newly generated layout scheme, and further optimize it through the conflict detection and resolution mechanism, and add the optimized layout scheme to the candidate set;

[0177] The layout evaluation and selection module 4 is used to comprehensively evaluate the candidate layout set using the Pareto optimization strategy and select the optimal layout solution as output;

[0178] The human-computer interaction module 5 is used to receive the layout space and furniture information input by the user, and present the optimized layout plan to the user.

[0179] Example 3

[0180] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in Example 1 when executing the computer program.

[0181] Example 4

[0182] A storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in Example 1 are implemented.

[0183] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, replacement, and improvement made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. The furniture layout method is characterized by: It includes the following steps: 1) Parse the layout space and furniture information input by the user to construct an initial layout plan; 2) According to the characteristics of the layout problem, use the adaptive window adjustment module to dynamically adjust the search window, and use this window to perform global or local optimization search on the basis of the current layout plan to generate a set of new layout plans; 3) Use the phase correction module to perform adaptive fine-tuning and correction on the newly generated layout plan, and further optimize the layout plan through the conflict detection and resolution mechanism, and add the optimized layout plan to the candidate set; 4) Repeat steps 2) and 3) until the iteration termination condition is met; 5) Use the Pareto optimization strategy to comprehensively evaluate the candidate layout set, and select the optimal layout plan as the output; In step 2), the adaptive window adjustment module adopts the Simulated Annealing-based Adaptive Window Search algorithm SAWS, which includes the following steps: 2.1) Initialize the size factor α and shape factor β of the search window, and the annealing temperature T, where α ∈ (0, 1), representing the size ratio of the search window to the entire layout space, and the initial value α0 = 0.5; β represents the degree of irregularity of the search window shape, and the initial value β0 = 0; the initial value T0 = 100; 2.2) Based on the current layout plan, use the search window to randomly select a group of furniture Perform local perturbations on its position and orientation to obtain a new local layout L′: Position perturbation: Within the search window, for each piece of furniture f i ∈F′, its position coordinates (x i ,y i ) is randomly offset with a step size of δ, that is, x i ′=x i +rand(-δ,δ),y i ′=y i +rand(-δ,δ), where rand(-δ,δ) means random sampling in the range of [-δ,δ]; Orientation perturbation: For each piece of furniture f i ∈F′, its orientation angle θ i Random deflection with step size η, that is, θ i ′=θ i +rand(-η,η); 2.3) Use the layout evaluation function E(·) to score the new layout L′, and decide whether to accept L′ as the new current layout according to the Metropolis criterion; if E(L′) < E(L), then accept L′ as the new current layout, otherwise, accept L′ as the new current layout with probability P = exp[-(E(L′) - E(L)) / T], where E(L) and E(L′) are the evaluation function values of the current layout and the new layout respectively; 2.4) According to the current number of iterations t and the maximum number of iterations T max , update the size factor α and shape factor β of the search window: α=α0×(1-tanh(k1×t / T max )) β=β0+(1-β0)×sin(k2×π×t / T max ) where k1 is the size adjustment rate factor and k2 is the shape adjustment rate factor; 2.5) According to the current number of iterations t and the maximum number of iterations T max , update the annealing temperature T T=T0×exp(-k3×t / T max ); where k3 is the cooling rate factor; 2.6) Repeat steps 2.2) to 2.5) until the iteration termination condition is met.

2. The method according to claim 1, characterized in that In step 3), the phase correction module adopts the Layout Phase Analysis-based Adaptive Correction algorithm LPAC, which includes the following steps: 3.1) Perform phase decomposition on the furniture layout in the candidate layout plan L to extract the following key phase components: Furniture spacing D = {d ij |i,j∈F,i≠j}, where d ij Represents the distance between furniture i and j, calculated as: d ij =[(x i -x j ) 2 +(y i -y j ) 2 ] 0.5 ; Alignment A = {a ij |i,j∈F,i≠j}, where a ij It indicates the alignment degree of furniture i and j in the x direction or y direction, and the calculation formula is: ij =min(|x i -x j |,|y i -y j |); Orientation consistency O = {o i |i∈F}, where o i represents the orientation angle θ of furniture i i The angle between the main direction of the room is calculated as: i =min(|θ i -θ0|,180°-|θ i -θ0|),θ0 is the main direction angle of the room; 3.2) Perform adaptive correction on each phase component to obtain the corrected layout L′: Spacing correction: For each pair of furniture (i, j), calculate the corrected spacing d i ' j =d ij +λ1·sign(d ij -d0)·log(1+|d ij -d0|), where d0 is the spacing reference value, λ1 is the spacing correction strength factor, and sign(·) is the sign function; Alignment correction: For each pair of furniture (i, j), calculate its corrected alignment a′ ij =a ij +λ2·sin(π·a ij / a0)·exp(-a ij / a0), where a0 is the alignment reference value and λ2 is the alignment correction strength factor; Orientation correction: For each piece of furniture i, calculate its corrected orientation angle θ′ i =θ i +λ3·tanh(o i / o0)·cos(2·o i ), where o0 is the orientation consistency reference value and λ3 is the orientation correction strength factor; 3.3) Reconstruct the layout L′ according to the corrected phase components and add it to the candidate set C as a new candidate layout.

3. The method according to claim 2, characterized in that After phase correction, use the constraint graph-based heuristic search algorithm to perform conflict detection and resolution on the layout plan, including the following steps: 4.1) Construct the layout constraint graph G = (V, E), where V is the set of furniture nodes and E is the set of constraint edges V = {v i |i∈F}, where v i Represents the node corresponding to furniture i; E={e ij |i,j∈F,i≠j,constraint(i,j)=True},where e i j represents node v i and v j The constraint edge between them, constraint (i, j) is a Boolean function indicating whether there is a layout constraint between furniture i and j; 4.2) Define the layout constraint rule set R, and common constraint rules include: Furniture does not overlap: Inter(i,j) = False, where Inter(i,j) is a Boolean function indicating whether furniture i and j have an intersection; Reasonable distance from the wall: d min ≤d(i,W)≤d max , where d(i,W) represents the distance from furniture i to the nearest wall, d min and d max is the lower and upper distance limit; Accessibility: i≠j, if there is a path from i to j, then Width(i,j)≥w minv Where Width(i,j) represents the minimum width of the path, w min is the lower limit of width; 4.3) Use the depth-first search algorithm to traverse the constraint graph G and mark the conflicting nodes and conflicting edges. Conflicting nodes: Make v i Violation of any rule r∈R; conflicting edge: i≠j, so that e ij Violation of any rule r∈R; 4.4) For each conflicting node v i, Perform heuristic adjustment according to the rule set R to obtain the adjusted furniture layout L′; if v i If the furniture does not overlap, a new position is randomly sampled in its neighborhood until it no longer overlaps. i If the "reasonable distance from the wall" rule is violated, move v in a direction perpendicular to the nearest wall. i, Until the distance limit is met; if v i If the "convenient passage" rule is violated, choose a suitable passage and adjust the position of related furniture so that the passage meets the width requirements; 4.5) For each conflicting edge e ij , perform heuristic adjustment according to the rule set R and obtain the adjusted furniture layout L′: if e ij If the "furniture does not overlap" rule is violated, adjust v separately. i and v j position until there is no overlap; if e ij If the "reasonable wall distance" rule is violated, adjust v i and v j positions so that they all satisfy the distance restriction; if e ij If the "traffic convenience" rule is violated, adjust v i and v j The location and layout of the passages make sure that the passages meet the width requirements; 4.6) Add the adjusted layout L′ to the candidate set C as a new candidate layout.

4. The method according to claim 1, characterized in that In step 4), the iteration termination condition includes reaching the maximum number of iterations T max ; The optimal layout score improvement of K consecutive iterations is less than the threshold ∈; The current temperature T is lower than the set minimum temperature T min .

5. The method according to claim 1 is characterized in that In step 5), the Pareto Optimization-based Multi-Objective Layout Evaluation and Selection algorithm PMOES is adopted, which includes the following steps: 6.1) Define a set of layout evaluation indicators U = {u1,u2,...,u n }, for each layout solution L in the candidate set C i , calculate its score on each indicator (u i1 ,u i2 ,...,u in ): Space utilization: u i1 =U si / S, where U si For layout L i The total area occupied by the furniture in the room, S is the total area of ​​the room; Lighting index: u i2 =W i / W, where W i For layout L i The total area of ​​all windows W is the total window area of ​​the room; Aesthetics: u i3 =(Sym i +Bal i +Div i ) / 3, where Sym i For layout L i Symmetry score, Bal i For layout L i Balance score of Div i For layout L i The richness score of , the score range is [0,1]; Comfort: where w j is the weight of the j-th comfort factor, g j is the actual value of the jth comfort factor, g j0 is the reference value of the jth comfort factor, and m is the total number of comfort factors; Personalization: i5 Given by users through questionnaires or preference feedback, reflecting the layout L i The degree of matching with the user's personalized needs, the value range is [0,1]; 6.2) Perform non-dominated sorting on the candidate set C based on the Pareto dominance relationship to obtain a set of Pareto fronts: Layout L i Pareto Dominant Layout L j′ If and only if u ik ≥u jk′ and u ik >u jk′ ;Layout L i With layout L j do not dominate each other if and only if u ik >u jk Andu il jl ; Pareto frontier ​ 6.3) Apply the weighted sum method on the Pareto frontier F1, combined with the user preference weight vector w = (w1, w2, ..., w n ), calculate each layout L i Overall ratings i : where w k is the weight of the kth indicator, w k ≥0; 6.4) Select the layout with the highest comprehensive score as the optimal recommended layout L * : 6.5) The optimal recommended layout L * It is returned together with other candidate layouts in the Pareto frontier F1 for user reference and selection; 6.6) Record the user's feedback evaluation of the recommended layout, dynamically update the preference weight vector w, and realize the continuous optimization of layout recommendation.

6. The method according to claim 1, characterized in that ,The layout space includes a room plan and a furniture list. The plan uses Canny edge detection and Hough ,transformation to extract the room boundary and the location of doors and windows. ,The furniture is represented parametrically, and the parameters include: center coordinates (x, y), length l, width w, and orientation θ.

7. A furniture layout device, implementing the method according to any one of claims 1 to 6, characterized in that , including: a layout parsing module, an adaptive window searching module, a layout phase correction module, a layout evaluation and selection module and a human-computer interaction module; The layout parsing module is used to parse the layout space and furniture information input by the user and construct an initial layout plan; The adaptive window search module is used to dynamically adjust the search window according to the characteristics of the layout problem, and use the window to perform global or local optimization search based on the current layout solution to generate a set of new layout solutions; The layout phase correction module is used to adaptively fine-tune and correct the newly generated layout scheme, and further optimize it through the conflict detection and resolution mechanism, and add the optimized layout scheme to the candidate set; The layout evaluation and selection module is used to comprehensively evaluate the candidate layout set using the Pareto optimization strategy and select the optimal layout solution as output; The human-computer interaction module is used to receive layout space and furniture information input by the user, and present the optimized layout plan to the user.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any method of claim 1-6 when executing the computer program.

9. A storage medium storing a computer program, wherein the computer program implements the steps of any method of claims 1 to 6 when executed by a processor.

Images