A maintenance scheduling method and system based on logical guidance and double-target ant colony

Abstract

The application discloses a kind of based on logic guidance and double target ant colony's maintenance scheduling method and system, the method includes: based on maintenance project analysis cluster data obtains discrete queue;Using iterative self-organizing data analysis algorithm to discrete queue iterative clustering obtains elevator team queue;The benchmark solution that elevator team queue satisfies hard constraint logic rule is calculated;To benchmark solution iterative addition soft constraint adjustment rule obtains initial scheduling strategy;Initial scheduling strategy is as hot start pheromone input scheduling optimization model;Management efficiency matrix and geographical efficiency matrix are set to scheduling optimization model and iteratively optimized;Scheduling optimization model outputs Pareto optimal solution set.Thereby, iterative self-organizing data analysis algorithm realizes maintenance resource balanced allocation, with initial scheduling strategy hot start, make convergence time reduce to second level, satisfy high concurrent scene under instant output demand, multidimensional constraint guidance mechanism has excellent global optimization performance, realizes the balanced allocation of manpower and maintenance work order.

Description

Technical Field

[0001] This invention relates to the field of scheduling planning technology, and in particular to a maintenance scheduling method and system based on logic guidance and dual-target ant colony. Background Technology

[0002] The elevator maintenance industry is characterized by strict cyclical nature, with each elevator typically requiring cleaning and maintenance every 15 days. Furthermore, elevator maintenance work is geographically dispersed, with elevators maintained by the same company often scattered across various locations. This necessitates that maintenance companies allocate human resources to manage the irregularly dispersed elevator projects within their jurisdiction within fixed scheduling cycles. This process involves an extremely complex set of constraints. On one hand, they must adhere to hard regulatory constraints (such as a maximum maintenance interval of 15 days, mandatory rest for night shifts, and user-specified maintenance dates), while on the other hand, they must also consider soft management constraints (such as prioritizing weekdays, ensuring a balanced number of elevators maintained, and addressing uneven geographical distribution).

[0003] Existing manual task scheduling methods typically rely on rigid grouping by geographical region and temporary staff allocation to address scheduling conflicts. This approach suffers from low scheduling efficiency, resulting in severe workload imbalances among maintenance personnel and deteriorating scheduling outcomes. Using greedy strategies or fixed rules (such as prioritizing projects nearing their expiration date or allocating staff based on proximity) for scheduling lacks consideration for the balance between maintenance distance and workload. Furthermore, due to the vast solution space, these algorithms face severe cold-start problems, with convergence times exceeding several hours. They are prone to getting trapped in local optima, lack global optimization capabilities, and fail to meet the high responsiveness requirements of scheduling operations. Summary of the Invention

[0004] The first aspect of this embodiment discloses a maintenance scheduling method based on logical guidance and dual-target ant colony, specifically including: Based on the analysis of cluster data from maintenance projects, discrete queues are obtained. An iterative self-organizing data analysis algorithm is used to iteratively cluster the discrete queue to obtain elevator team columns; Calculate the baseline solution for the elevator team column that satisfies the hard constraint logic rules; Soft constraint adjustment rules are added iteratively to the benchmark solution to obtain the initial scheduling strategy; The initial scheduling strategy is used as a hot-start pheromone input to the scheduling optimization model; For the aforementioned scheduling optimization model, a management efficiency matrix and a geographic efficiency matrix are set and iteratively optimized; The scheduling optimization model outputs a Pareto optimal solution set.

[0005] As an optional implementation, the method further includes: Calculate the gradient density of the stationed project; Stationary projects with a gradient density exceeding the density threshold or a coverage radius exceeding the radius threshold are set as cluster data.

[0006] As an optional implementation, the step of using an iterative self-organizing data analysis algorithm to iteratively cluster the discrete queue to obtain elevator team columns includes: Initialize and generate the initial center point; Calculate the distance from each maintenance item to each initial center point, and assign each maintenance item to the corresponding temporary area based on the principle of minimum distance. Set a spatial constraint condition that the distance between any two maintenance projects in the same temporary area must not exceed a distance threshold; Maintenance items that exceed the spatial constraints are reassigned to other temporary areas of the next best option until all maintenance items meet the spatial constraints. Then, the divided temporary areas are aggregated to form the elevator team.

[0007] As an optional implementation, the method further includes: If the total number of elevators or the maximum standard deviation exceeds the limit in any temporary area, then that temporary area will be divided into two temporary areas. If the distance between the centers of any two temporary regions is less than the merging threshold, they are then merged into one temporary region by weighted average. For each temporary region, the contour coefficient and DB index are monitored. If the contour coefficient and DB index do not improve for 5 consecutive rounds, the iteration is terminated.

[0008] As an optional implementation, calculating the baseline solution for the elevator team column to satisfy hard constraint logic rules includes: Read the oldest last maintenance time from the elevator team column and map it to the current shift cycle; For subsequent elevator groups, a greedy strategy is adopted, inserting the nearest and not fully loaded shift time in the current shift cycle; By performing local scans, scheduling times that violate the hard constraint logic rules are corrected cyclically.

[0009] As an optional implementation, the step of iteratively adding soft constraint adjustment rules to the benchmark solution to obtain the initial scheduling strategy includes: The smallest adjustable unit in the baseline solution is adjusted based on the soft constraint adjustment rule, and the loss is calculated after the adjustment is completed. If the loss is reduced, then based on the current version of the scheduling strategy, the next smallest adjustable unit will be adjusted; If losses increase, revert to the previous version of the scheduling strategy and adjust the next minimum adjustable unit. When the cycle is adjusted until the loss no longer decreases or the maximum number of iterations is reached, the initial scheduling strategy is obtained through iteration.

[0010] As an optional implementation, the step of setting a management efficiency matrix and a geographic efficiency matrix for the scheduling optimization model and iteratively optimizing them includes: The management efficiency matrix and the geographic efficiency matrix are assigned a pheromone concentration corresponding to 10 times the initial pheromone concentration of the baseline solution path; The management efficiency matrix aggregation includes soft constraint adjustment rules such as workday priority, daily task volume limit, task balance and the number of scattered ladders, in order to improve the rationality of human resource allocation; The geographic performance matrix aggregation includes soft constraint adjustment rules with staggered ladder distance limits and geographic proximity to improve the rationality of commuting scheduling; In a single iteration, the pheromone concentrations of the management efficiency matrix and the geographic efficiency matrix are read simultaneously, the comprehensive selection probability is calculated to obtain the shortest path solution, and it is also calculated whether the hard constraint logic rule is violated.

[0011] As an optional implementation, the scheduling optimization model outputs a Pareto optimal solution set, including: For each new solution generated, determine whether it is dominated by any historical output solution; If the new solution is not assigned, it is archived. If the number of new solutions output reaches the archive limit, redundant solutions are removed based on the crowding level of each solution in the archive. When the maximum number of iterations is reached or the scheduling optimization model converges, the archived output will be the Pareto optimal solution set.

[0012] As an optional implementation, the method further includes: Based on the Pareto optimal solution set, the management efficiency matrix and the geographical efficiency matrix are updated in reverse.

[0013] The second aspect of this embodiment discloses a maintenance scheduling system based on logical guidance and dual-target ant colony, specifically including: The filtering unit is used to analyze cluster data based on maintenance projects to obtain discrete queues; Clustering units are used to iteratively cluster the discrete queue using an iterative self-organizing data analysis algorithm to obtain elevator team queues; Hard constraint unit, used to calculate the benchmark solution of the elevator team column satisfying the hard constraint logic rules; A soft constraint unit is used to iteratively add soft constraint adjustment rules to the benchmark solution to obtain an initial scheduling strategy; A hot-start unit is used to input the initial scheduling strategy as a hot-start pheromone into the scheduling optimization model; An iterative unit is used to set the management efficiency matrix and the geographic efficiency matrix for the scheduling optimization model and to iteratively optimize them; The output unit is used to output the Pareto optimal solution set.

[0014] Compared with the prior art, this embodiment has the following beneficial effects: The system achieves balanced allocation of maintenance resources based on iterative self-organizing data analysis algorithm. It also uses an initial scheduling strategy to start the model, reducing the convergence time from hours to seconds, meeting the real-time output requirements in high-concurrency scenarios. The multi-dimensional constraint guidance mechanism has excellent global optimization performance, realizing balanced allocation of manpower and maintenance work orders. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this embodiment, the accompanying drawings used in the embodiment will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the workflow of a maintenance scheduling method based on logic guidance and dual-target ant colony disclosed in this embodiment; Figure 2 This is a schematic diagram of the system structure of a maintenance scheduling system based on logic guidance and dual-target ant colony disclosed in this embodiment. Detailed Implementation

[0017] The technical solutions in this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1 Please see Figure 1 This embodiment discloses a maintenance scheduling method based on logical guidance and dual-target ant colony, including: 101. Based on the maintenance project analysis cluster data, obtain discrete queues.

[0019] In this embodiment, the ladder density of the stationed project is calculated; Stationary projects with a gradient density exceeding the density threshold or a coverage radius exceeding the radius threshold are set as cluster data.

[0020] It is understandable that, because some on-site projects have a large number of elevator maintenance work orders in a very small area, the location of the on-site project appears as a significant outlier in the overall data. If it is included in the overall optimization calculation, other discrete work orders will be significantly affected, resulting in an unsatisfactory overall optimization effect.

[0021] Therefore, instead of using a single elevator as the dispatch unit, we use elevator clusters as the smallest courtyard unit to strongly bind elevators that are maintained on the same day in the same project.

[0022] 102. An iterative self-organizing data analysis algorithm is used to iteratively cluster discrete queues to obtain elevator team queues.

[0023] In this embodiment, elevator clusters can not only more reasonably represent the actual distance of maintenance work orders in a geographical area, but also significantly reduce the number of terminals in the optimization calculation process (e.g., 1,000 elevators are reduced to 200 elevator clusters), thereby improving the efficiency of optimization calculation.

[0024] As an optional implementation, an iterative self-organizing data analysis algorithm is used to iteratively cluster the discrete queues to obtain elevator team columns, including: Initialize and generate the initial center point; Calculate the distance from each maintenance item to each initial center point, and assign each maintenance item to the corresponding temporary area based on the principle of minimum distance. Set a spatial constraint condition that the distance between any two maintenance projects in the same temporary area must not exceed a distance threshold; Maintenance projects that exceed the spatial constraints are reassigned to other temporary areas of the next best choice until all maintenance projects meet the spatial constraints. Then, the divided temporary areas are aggregated to form an elevator team.

[0025] As an optional implementation, if the total number of elevators or the maximum standard deviation in any temporary area exceeds the limit, the temporary area is divided into two temporary areas. If the distance between the centers of any two temporary regions is less than the merging threshold, they are then merged into one temporary region by weighted average. For each temporary region, the contour coefficient and DB index are monitored. If the contour coefficient and DB index do not improve for 5 consecutive rounds, the iteration is terminated.

[0026] Specifically, the distance between each other is used as a key constraint to minimize the fragmented travel distances between different maintenance work orders.

[0027] 103. Calculate the baseline solution for the elevator team to satisfy the hard constraint logic rules.

[0028] In this embodiment, hard constraint logic rules are used as the limiting conditions for subsequent optimization algorithms to avoid local optimization or optimization results that do not meet actual requirements.

[0029] As an optional implementation, the benchmark solution for calculating the elevator team column to satisfy the hard constraint logic rules includes: Read the oldest last maintenance time from the elevator team column and map it to the current shift cycle; For subsequent elevator groups, a greedy strategy is adopted, inserting the nearest and not fully loaded shift time in the current shift cycle; By performing local scans, the scheduling times that violate hard constraint logic rules are corrected in a loop.

[0030] Here, the longest time since the last maintenance is taken as the limit, and the maximum maintenance interval is strictly guaranteed. For subsequent elevator groups whose maintenance period has not yet approached, a greedy strategy is adopted to insert the nearest and not fully loaded shift time. At the same time, the logic rules that violate the hard constraints such as the specified workday and the night shift rest interval are cyclically corrected to ensure that the benchmark solution is reasonable and feasible.

[0031] 104. Add soft constraint adjustment rules to the benchmark solution iteratively to obtain the initial scheduling strategy.

[0032] In this embodiment, the soft constraint adjustment rule is used to optimize the scheduling based on the baseline solution, so as to achieve the maximum benefit of both management efficiency and geographical efficiency.

[0033] As an optional implementation, soft constraint adjustment rules are added to the benchmark solution iteration to obtain the initial scheduling strategy, including: The smallest adjustable unit in the baseline solution is adjusted based on the soft constraint adjustment rule, and the loss is calculated after the adjustment is completed. If the loss is reduced, then based on the current version of the scheduling strategy, the next smallest adjustable unit will be adjusted; If losses increase, revert to the previous version of the scheduling strategy and adjust the next minimum adjustable unit. When the loop is adjusted until the loss no longer decreases or the maximum number of iterations is reached, the iteration obtains the initial scheduling strategy.

[0034] Here, under the adjustment of the soft constraint adjustment rules, the work order scheduling that is significantly concentrated in the baseline scheduling strategy will be locally optimized.

[0035] 105. The initial scheduling strategy is used as the hot-start pheromone input for the scheduling optimization model.

[0036] In this embodiment, starting the scheduling optimization model based on the initial scheduling strategy can ensure that the solution space is reduced by an order of magnitude, and ensure that the scheduling optimization model can be successfully started without wasting a lot of computing power and time in the ineffective or inefficient convergence process.

[0037] 106. Set up management efficiency matrix and geographic efficiency matrix for the scheduling optimization model and iteratively optimize them.

[0038] In this embodiment, by constructing two independently evolving matrices, for two types of mutually conflicting optimization objectives, it is ensured that the output term will not undergo local optimization.

[0039] As an optional implementation, a management performance matrix and a geographic performance matrix are set for the scheduling optimization model and iteratively optimized, including: The management efficiency matrix and the geographic efficiency matrix are assigned a pheromone concentration that is 10 times the initial pheromone concentration of the baseline solution path; The management efficiency matrix aggregation includes soft constraint adjustment rules such as workday priority, daily task volume limit, task balance and the number of scattered ladders, in order to improve the rationality of human resource allocation; The geographic performance matrix aggregation includes soft constraint adjustment rules with staggered ladder distance limits and geographic proximity to improve the rationality of commuting scheduling; In a single iteration, the pheromone concentrations of the management efficiency matrix and the geographic efficiency matrix are read simultaneously to calculate the comprehensive selection probability in order to obtain the shortest path solution, while also calculating whether the hard constraint logic rules are violated.

[0040] Here, the management efficiency matrix and the geographical efficiency matrix mutually influence each other's outputs during the optimization process, achieving a balanced optimization.

[0041] 107. The scheduling optimization model outputs the Pareto optimal solution set.

[0042] In this embodiment, the scheduling optimization model does not retain only a single optimal solution, but continuously maintains a finite-capacity archive to accommodate non-dominated solutions (options that are not inferior to other solutions in all objective dimensions), thereby guiding the continuous approximation of the Pareto front to find a balance point that is "both balanced and close".

[0043] As an optional implementation, the scheduling optimization model outputs a Pareto optimal solution set, including: For each new solution generated, determine whether it is dominated by any historical output solution; If the new solution is not assigned, it is archived. If the number of new solutions output reaches the archive limit, redundant solutions are removed based on the crowding level of each solution in the archive. When the maximum number of iterations is reached or the scheduling optimization model converges, the archived output will be a Pareto optimal solution set.

[0044] As an optional implementation, the management performance matrix and the geographic performance matrix are updated in reverse based on the Pareto optimal solution set.

[0045] Here, the elite solutions in the archive are also used to update the performance matrix in reverse, making both more likely to find the optimal solution in the subsequent optimization process and improving convergence efficiency.

[0046] Here, the nearest target is prioritized as the valid evidence for this verification, and this is used to match the shooting perspective of the maintenance personnel.

[0047] Compared with the prior art, this embodiment has the following beneficial effects: The system achieves balanced allocation of maintenance resources based on iterative self-organizing data analysis algorithm. It also uses an initial scheduling strategy to start the model, reducing the convergence time from hours to seconds, meeting the real-time output requirements in high-concurrency scenarios. The multi-dimensional constraint guidance mechanism has excellent global optimization performance, realizing balanced allocation of manpower and maintenance work orders.

[0048] Example 2 Please see Figure 2 This embodiment discloses a maintenance scheduling system based on logic guidance and dual-target ant colony, including: The filtering unit is used to analyze cluster data based on maintenance projects to obtain discrete queues; Clustering units are used to iteratively cluster discrete queues using an iterative self-organizing data analysis algorithm to obtain elevator team queues; Hard constraint cells are used to calculate the baseline solution for elevator group columns to satisfy hard constraint logic rules; Soft constraint units are used to iteratively add soft constraint adjustment rules to the baseline solution to obtain the initial scheduling strategy; The hot start unit is used to input the initial scheduling strategy as a hot start pheromone into the scheduling optimization model; The iterative unit is used to set the management efficiency matrix and geographic efficiency matrix for the scheduling optimization model and to iteratively optimize them; The output unit is used to output the Pareto optimal solution set.

Claims

1. A maintenance scheduling method based on logical guidance and dual-target ant colony, characterized in that, include: Based on the analysis of cluster data from maintenance projects, discrete queues are obtained. An iterative self-organizing data analysis algorithm is used to iteratively cluster the discrete queue to obtain elevator team columns; Calculate the baseline solution for the elevator team column that satisfies the hard constraint logic rules; Soft constraint adjustment rules are added iteratively to the benchmark solution to obtain the initial scheduling strategy; The initial scheduling strategy is used as a hot-start pheromone input to the scheduling optimization model; For the aforementioned scheduling optimization model, a management efficiency matrix and a geographic efficiency matrix are set and iteratively optimized; The scheduling optimization model outputs a Pareto optimal solution set.

2. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 1, characterized in that, The method further includes: Calculate the gradient density of the stationed project; Stationary projects with a gradient density exceeding the density threshold or a coverage radius exceeding the radius threshold are set as cluster data.

3. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 1, characterized in that, The step of using an iterative self-organizing data analysis algorithm to iteratively cluster the discrete queue to obtain elevator team columns includes: Initialize and generate the initial center point; Calculate the distance from each maintenance item to each initial center point, and assign each maintenance item to the corresponding temporary area based on the principle of minimum distance. Set a spatial constraint condition that the distance between any two maintenance projects in the same temporary area must not exceed a distance threshold; Maintenance items that exceed the spatial constraints are reassigned to other temporary areas of the next best option until all maintenance items meet the spatial constraints. Then, the divided temporary areas are aggregated to form the elevator team.

4. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 3, characterized in that, The method further includes: If the total number of elevators or the maximum standard deviation exceeds the limit in any temporary area, then that temporary area will be divided into two temporary areas. If the distance between the centers of any two temporary regions is less than the merging threshold, they are then merged into one temporary region by weighted average. For each temporary region, the contour coefficient and DB index are monitored. If the contour coefficient and DB index do not improve for 5 consecutive rounds, the iteration is terminated.

5. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 1, characterized in that, The calculation of the baseline solution for the elevator team column satisfying the hard constraint logic rules includes: Read the oldest last maintenance time from the elevator team column and map it to the current shift cycle; For subsequent elevator groups, a greedy strategy is adopted, inserting the nearest and not fully loaded shift time in the current shift cycle; By performing local scans, scheduling times that violate the hard constraint logic rules are corrected cyclically.

6. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 1, characterized in that, The step of iteratively adding soft constraint adjustment rules to the benchmark solution to obtain the initial scheduling strategy includes: The smallest adjustable unit in the baseline solution is adjusted based on the soft constraint adjustment rule, and the loss is calculated after the adjustment is completed. If the loss is reduced, then based on the current version of the scheduling strategy, the next smallest adjustable unit will be adjusted; If losses increase, revert to the previous version of the scheduling strategy and adjust the next minimum adjustable unit. When the cycle is adjusted until the loss no longer decreases or the maximum number of iterations is reached, the initial scheduling strategy is obtained through iteration.

7. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 1, characterized in that, The step of setting and iteratively optimizing the management efficiency matrix and geographical efficiency matrix for the scheduling optimization model includes: The management efficiency matrix and the geographic efficiency matrix are assigned a pheromone concentration corresponding to 10 times the initial pheromone concentration of the baseline solution path; The management efficiency matrix aggregation includes soft constraint adjustment rules such as workday priority, daily task volume limit, task balance and the number of scattered ladders, in order to improve the rationality of human resource allocation; The geographic performance matrix aggregation includes soft constraint adjustment rules with staggered ladder distance limits and geographic proximity to improve the rationality of commuting scheduling; In a single iteration, the pheromone concentrations of the management efficiency matrix and the geographic efficiency matrix are read simultaneously, the comprehensive selection probability is calculated to obtain the shortest path solution, and it is also calculated whether the hard constraint logic rule is violated.

8. The maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 7, characterized in that, The scheduling optimization model outputs a Pareto optimal solution set, including: For each new solution generated, determine whether it is dominated by any historical output solution; If the new solution is not assigned, it is archived. If the number of new solutions output reaches the archive limit, redundant solutions are removed based on the crowding level of each solution in the archive. When the maximum number of iterations is reached or the scheduling optimization model converges, the archived output will be the Pareto optimal solution set.

9. A maintenance scheduling method based on logical guidance and dual-target ant colony as described in claim 8, characterized in that, The method further includes: Based on the Pareto optimal solution set, the management efficiency matrix and the geographical efficiency matrix are updated in reverse.

10. A maintenance scheduling system based on logic guidance and dual-target ant colony, characterized in that, include: The filtering unit is used to analyze cluster data based on maintenance projects to obtain discrete queues; Clustering units are used to iteratively cluster the discrete queue using an iterative self-organizing data analysis algorithm to obtain elevator team queues; Hard constraint unit, used to calculate the benchmark solution of the elevator team column satisfying the hard constraint logic rules; A soft constraint unit is used to iteratively add soft constraint adjustment rules based on maintenance project analysis cluster data to the benchmark solution to obtain an initial scheduling strategy. A hot-start unit is used to input the initial scheduling strategy as a hot-start pheromone into the scheduling optimization model; An iterative unit is used to set the management efficiency matrix and the geographic efficiency matrix for the scheduling optimization model and to iteratively optimize them; The output unit is used to output the Pareto optimal solution set.

Images