A mobile drainage pump truck optimization scheduling method based on pumping time sequence embedding

Abstract

The application discloses a kind of mobile flood drainage pump truck optimization scheduling method based on pumping timing embedding, it is related to urban mobile flood drainage equipment scheduling technical field, constructs one-dimensional drainage pipe network-two-dimensional surface overland flow coupling model, collects target area history and simulated waterlogging position, forms candidate operating position set after screening;Randomly generate each pump truck initial operating position, generate whole process operation schedule in combination with time, speed and accessibility constraint, integrate multiple vehicle schedule into optimization individual, optimize as a whole with minimum comprehensive waterlogging influence as target. Operation schedule is converted into corresponding two-dimensional grid time-varying negative flow sequence, embedded hydrodynamic calculation is carried out through one-dimensional node interface, update waterlogging process in combination with grid available water amount, and elite reservation genetic algorithm is used to iteratively solve optimal scheduling plan, the closed loop of complete operation schedule, pumping timing embedding, hydrodynamic simulation and flood drainage effect evaluation is realized by the application, and the overall regional flood drainage efficiency under the condition of limited pump truck can be improved.

Description

Technical Field

[0001] This invention relates to the field of urban mobile drainage equipment scheduling technology, and more specifically to an optimized scheduling method for mobile drainage pump trucks based on pumping timing embedding. Background Technology

[0002] Currently, under conditions of short-duration heavy rainfall, continuous torrential rain, or overloaded urban fixed drainage facilities, roads, underpasses, residential areas, and areas surrounding important public facilities are prone to flooding. Mobile drainage pumping vehicles, with their flexible deployment, high pumping capacity, and ability to move between multiple flooded areas, are essential equipment for urban flood emergency response.

[0003] Existing pump truck deployments are typically determined based on historical water accumulation points, field experience, or current water depth. When multiple water accumulation points occur simultaneously and the number of pump trucks is limited, selecting only the deepest or closest location makes it difficult to guarantee the effectiveness of multiple pump trucks in reducing the total water volume in the area.

[0004] Existing methods often select the next work point sequentially based on the current water accumulation status, making local decisions at each moment. Such point-by-point or rolling methods make it difficult to compare the initial location, subsequent location sequence, work time at each location, and movement time within an emergency cycle as a complete plan, and also make it difficult to evaluate the combined drainage effect of multiple pump trucks in a complete plan.

[0005] When evaluating pump truck scheduling schemes using hydrodynamic models, it is also necessary to address how to incorporate pump truck location and pumping flow rate into two-dimensional water accumulation calculations. Two-dimensional surface models typically receive water volume input through one-dimensional node flow time series, while the pump truck location changes over time. Simply subtracting water volume from the model's external data is insufficient to reflect the continuous feedback from inter-grid confluence, one-dimensional to two-dimensional exchange, and pumping action.

[0006] Therefore, how to provide a method that can generate a complete operation plan table for multiple mobile pumping trucks throughout the entire process, convert the complete operation plan table into a pumping time series that can be read by the model, and optimize the complete operation plan of all pumping trucks as a whole with the goal of minimizing the combined impact of water accumulation area, water accumulation depth and water accumulation duration, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides an optimized scheduling method for mobile drainage pumping trucks based on pumping sequence embedding, which aims to solve the problems in the prior art where the initial position of the pumping truck and subsequent operation arrangements rely on local experience, the complete operation plan is difficult to optimize as a whole, the pumping truck scheduling scheme is difficult to embed into two-dimensional hydrodynamic calculations, and the optimization evaluation index cannot fully reflect the water accumulation range, depth and duration.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: An optimized scheduling method for mobile drainage pumping vehicles based on pumping sequence embedding includes: Acquire basic data of the target area, and construct an urban waterlogging hydrological and hydrodynamic model based on the basic data, including a one-dimensional drainage network model and a two-dimensional surface runoff model. The historical actual water accumulation locations in the target area are summarized with the water accumulation locations simulated by the urban waterlogging hydrological and hydrodynamic model. A candidate operation location set is formed through preliminary manual screening. A mapping relationship is established between the two-dimensional grid numbers in the candidate operation location set and the one-dimensional model sampling input node numbers. The initial working position of each pump truck is randomly generated from the candidate working position set, and a complete work plan table containing subsequent working positions, working duration and moving time is generated based on the remaining working time, pump truck moving speed, minimum working time and location reachability constraints. Combine the complete work schedules of all pump trucks into a single scheduling scheme to be optimized. According to the complete work plan, the rated drainage flow for each work period is converted into a time-varying negative flow sequence for the corresponding drainage input node, and the time-varying negative flow sequence is input into the two-dimensional surface runoff model. At each calculation time step, the actual pumping volume is determined based on the available water volume in the current two-dimensional grid, the water depth in the two-dimensional grid is updated, and the target value of the single comprehensive water accumulation impact is calculated. A single-objective optimization algorithm is adopted, and the complete work plan table is used as the optimization individual for iterative updates. The mobile drainage pumping truck full-process scheduling plan corresponding to the minimum comprehensive water accumulation impact target value is output.

[0009] Optionally, the two-dimensional surface runoff model uses the node flow time series in the one-dimensional drainage network model as the water input interface; the two-dimensional grid number corresponds one-to-one with the pumping input node number, and the pumping input node is used to carry the pumping flow time series of the mobile drainage pump truck at the corresponding two-dimensional grid.

[0010] Optionally, the one-dimensional drainage network model is constructed using SWMM, and the pumping input node is implemented by adding manhole nodes corresponding to the candidate two-dimensional grids in the SWMM model. When the pump truck is dispatched to the target two-dimensional grid, the pumping flow time series of the corresponding manhole node is activated, and the two-dimensional surface runoff model reads the pumping flow time series and applies the corresponding flow to the target two-dimensional grid. Optionally, the candidate operation location set is formed by manually screening the historical actual water accumulation locations and the water accumulation locations simulated by the hydrological and hydrodynamic model in the target area. The manual preliminary screening is used to eliminate locations that are obviously unsuitable as operation points for mobile drainage pump trucks, and the screening results are saved through a candidate grid tag file. Optionally, the complete operation plan table for each pump truck covering the preset total operation cycle includes: multiple operation grids arranged in an orderly manner, the operation duration in each operation grid, and the movement time between adjacent operation grids. The first operation location is randomly generated from the candidate operation location set, and subsequent operation locations are generated from candidate locations that meet the remaining time and movement accessibility constraints. The operation duration and movement time together cover the preset total operation cycle.

[0011] Optionally, the scheduling scheme for all pump trucks is composed of the work plans of each pump truck. The arrival time of each work location is recursively derived based on the arrival time, work time, and travel time of the previous location. Simultaneously, the work completion time for a given work location is calculated. Each pump truck starts from a randomly generated first work location and, combined with the remaining work time, travel speed, and reachable range, sequentially generates subsequent work locations, work durations, and travel times until a complete work plan covering the entire work cycle is formed. All the complete work plans for the pump trucks together constitute an optimized unit, and the first work location, subsequent locations, and related time parameters are evaluated and updated holistically within the same optimized unit. Optionally, the conversion of the time-varying negative flow sequence specifically involves: The two-dimensional grid corresponding to each pump truck in each operating period is directly read from the complete work plan. During the pump truck's operating period, the flow rate of the corresponding pumping input node in the two-dimensional grid is set to a negative rated pumping flow rate of the pump truck. During the pump truck's moving period, the flow rate of the relevant pumping input node is set to 0. When multiple pump trucks are allowed to operate simultaneously in the same grid, the rated pumping flow rates of each pump truck are superimposed. Optionally, the calculation of the single comprehensive water accumulation impact target value specifically involves: First, water depth intervals corresponding to light, moderate, and severe water accumulation are defined. A level indicator function is used to determine the water accumulation level of each grid at each calculation time, and the cumulative duration of each grid corresponding to different water accumulation levels is statistically obtained. Then, based on the set three-level water accumulation weights, the cumulative duration of each level within a single grid is weighted and summed to obtain the weighted water accumulation impact of a single grid. Finally, the weighted water accumulation impact of all grids is multiplied by the corresponding grid area and summed to form a single comprehensive water accumulation impact target value that simultaneously reflects the water accumulation range, water accumulation depth, and water accumulation duration. Optionally, the single-objective optimization algorithm is a genetic algorithm based on an elite preservation strategy. Multiple scheduling schemes that satisfy the constraints of candidate location, total operation cycle, minimum operation time, and mobility reachability are randomly generated and form an initial population. Each scheduling scheme contains a complete operation plan table for K pump trucks. Subsequently, the optimization algorithm takes the overall impact of water accumulation as the single objective and finds the scheme with the minimum overall impact of water accumulation in the set of all feasible complete operation plans. The generation of the next operation location is only used to construct a complete plan table that satisfies the time and mobility constraints. The comparison and update of the optimization algorithm are carried out on the entire complete operation plan table, rather than optimizing the next operation location one by one during the operation.

[0012] Optionally, the mobile drainage pump truck's full-process scheduling plan includes at least the pump truck number, initial work grid number, each subsequent work grid number, arrival time, work start time, work end time, movement time, set drainage flow rate, actual effective drainage flow rate, and the corresponding drainage flow rate time series file.

[0013] As can be seen from the above technical solution, compared with the prior art, the present invention discloses an optimized scheduling method for mobile drainage pumping trucks based on pumping and drainage timing embedding, which has the following beneficial effects: 1. This invention aims to optimize pump truck scheduling schemes and uses pumping and drainage timing embedding as a model calculation method to form a closed loop of scheduling scheme generation, hydrodynamic simulation, water accumulation impact assessment, and scheme iterative update.

[0014] 2. This invention combines historical actual water accumulation locations with model-simulated water accumulation locations, and forms a set of candidate operation locations through preliminary manual screening. This can reduce the optimization search space and reduce invalid iterations while retaining real-world water accumulation experience and potential risk locations.

[0015] 3. This invention encodes the initial working position, subsequent position sequence, working time and movement time of each pump truck into a complete work plan table, and compares all the complete work plans of multiple pump trucks as an optimized individual, avoiding local decision-making caused by selecting the next working position one by one based solely on the current state.

[0016] 4. This invention converts the two-dimensional grid where the pump truck is located, the operating period, and the pumping flow rate into the negative flow time series of the corresponding pumping input node. The two-dimensional calculation is embedded through the existing one-dimensional node time series interface, without the need to directly construct a continuously moving entity in the two-dimensional model.

[0017] 5. This invention employs a comprehensive objective that simultaneously reflects the water accumulation range, water accumulation depth, and water accumulation duration, thereby improving the characterization ability of the overall regional drainage effect while maintaining the optimized structure of a single objective.

[0018] 6. This invention can output a full-process scheduling plan including the initial position of the pump truck, each working position, arrival time, working time, movement time and pumping flow rate, and simultaneously generate a pumping time sequence file that can be used for model recalculation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 A schematic diagram illustrating the one-dimensional-two-dimensional coupled calculation and pumping timing embedding principle of the mobile drainage pumping vehicle provided by the present invention; Figure 2 This is a schematic diagram of the overall process of the mobile drainage pumping vehicle optimization scheduling method provided by the present invention; Figure 3 This invention provides a schematic diagram illustrating the mapping relationship between one-dimensional model sampling input nodes and two-dimensional computational grid numbers. Figure 4 A schematic diagram illustrating the complete operation plan coding and constraints of the mobile drainage pump truck provided by this invention; Figure 5 This is a schematic diagram illustrating the conversion of a complete job plan into a node negative flow time series and an example of model input provided by the present invention. Figure 6 This is a schematic diagram illustrating the optimization solution process for the mobile drainage pump truck scheduling scheme based on genetic algorithms provided by the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention 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.

[0022] This invention discloses an optimized scheduling method for mobile drainage pumping vehicles based on pumping time sequence embedding, such as... Figure 1 and Figure 2 As shown, it includes: Acquire basic data of the target area, and construct an urban waterlogging hydrological and hydrodynamic model based on the basic data, including a one-dimensional drainage network model and a two-dimensional surface runoff model. The historical actual water accumulation locations in the target area are summarized with the water accumulation locations simulated by the urban waterlogging hydrological and hydrodynamic model. A candidate operation location set is formed through preliminary manual screening. A mapping relationship is established between the two-dimensional grid numbers in the candidate operation location set and the one-dimensional model sampling input node numbers. The initial working position of each pump truck is randomly generated from the candidate working position set, and a complete work plan table containing subsequent working positions, working duration and moving time is generated based on the remaining working time, pump truck moving speed, minimum working time and location reachability constraints. Combine the complete work schedules of all pump trucks into a single scheduling scheme to be optimized. According to the complete work plan, the rated drainage flow for each work period is converted into a time-varying negative flow sequence for the corresponding drainage input node, and the time-varying negative flow sequence is input into the two-dimensional surface runoff model. At each calculation time step, the actual pumping volume is determined based on the available water volume in the current two-dimensional grid, the water depth in the two-dimensional grid is updated, and the target value of the single comprehensive water accumulation impact is calculated. A single-objective optimization algorithm is adopted, and the complete work plan table is used as the optimization individual for iterative updates. The mobile drainage pumping truck full-process scheduling plan corresponding to the minimum comprehensive water accumulation impact target value is output.

[0023] The following section details the specific content of each part.

[0024] ① Two-dimensional grid and set of candidate job locations

[0025] The two-dimensional surface model of the target area is divided into M computational grids. Each grid has a unique number and records the row and column number, center coordinates, and grid area. These attributes are used for candidate location selection, distance calculation, and water accumulation impact statistics, and are no longer considered as separate optimization variables.

[0026] Summary of historical actual water accumulation locations in the target area S h And the set of locations of water accumulation simulated by hydrological and hydrodynamic models S m Researchers conducted preliminary screening based on their practical experience in flood drainage, forming a set of candidate operation locations. S : (1) In equation (1), S h This is a collection of historical actual water accumulation locations. S m For hydrodynamic models to simulate the set of water accumulation locations, F This indicates a preliminary manual screening operation. This preliminary manual screening is used to eliminate locations that are clearly unsuitable as operating sites for mobile drainage pumping vehicles, and the screening results can be saved in a candidate grid marker file. This step is completed before the optimization iteration and is not considered an optimization variable; it is used to narrow the location search space, reduce the evaluation of invalid solutions, and improve the practical feasibility of the scheduling plan.

[0027] ② Timing mapping between two-dimensional mesh and sampling input nodes

[0028] like Figure 3 As shown, corresponding sampling input nodes are established for the two-dimensional grids in the candidate job location set. The mapping relationship between the two-dimensional grid number and the sampling input node number is as follows: (2) In equation (2), ψ For mapping functions, It is a two-dimensional grid. N i This corresponds to the pumping input nodes. In one specific implementation, the one-dimensional drainage network model uses SWMM, and the two-dimensional model uses the one-dimensional manhole node flow time series as the input source; for candidate two-dimensional grids... g i Add a number to the SWMM model. N i The inspection well node. This node serves as the data interface for the pump truck's pumping flow rate to enter the two-dimensional model.

[0029] The pump truck remained on the two-dimensional grid. g i When, enable node N i Negative flow time series; after the pump truck leaves the grid. N i The flow rate was restored to 0. By activating the node time series corresponding to different grids at different times, the pump truck's pumping action was embedded in two-dimensional space and time.

[0030] ③ Generation of complete work plan and overall coding

[0031] like Figure 4 As shown, suppose K mobile drainage pumping vehicles are deployed. When generating a scheduling scheme to be evaluated, the first operating position of each pumping vehicle is randomly selected from the candidate operating position set S; in the implementation method that does not allow initial position overlap, random sampling without replacement is used: (3) In equation (3), This is the first operating position of the k-th pump truck. Sample(S,K) This indicates that K distinct locations are randomly selected from the candidate job location set S. If multiple pump trucks are allowed to have the same initial location, random sampling with replacement can be used.

[0032] The work plan for the kth pump truck during the entire scheduling cycle is represented as follows: (4) In equation (4), For the first r Each work location For the duration of the operation at that location, The time taken to move from this position to the next position. This represents the number of operating locations for the pump truck. The scheduling plan for all pump trucks is as follows: (5) The arrival time of each work location is calculated by recursively calculating the arrival time of the previous location, the work time, and the travel time: (6) No. r The completion time for each job location is: (7) For each pump truck, starting from the first randomly generated work location, subsequent work locations, work durations, and travel times are generated sequentially based on the remaining work time, movement speed, and reachability, until a complete work plan covering the entire work cycle is formed. The complete work plans for all K pump trucks together constitute an optimization unit; the first work location, subsequent locations, and time parameters are evaluated and updated as a whole within the same optimization unit. The travel time, reachability, and other scheduling constraints involved in generating the complete work plan will be further explained later.

[0033] ④ Generate the pumping flow time series based on the complete work plan.

[0034] like Figure 5 As shown, the pump truck's operating location, start time, end time, and relocation time are all recorded in the complete work plan table. The model no longer determines the pump truck's location separately; instead, it reads the work grid and work time periods segment by segment from the plan table. k pump truck r The traffic time series for each work period corresponding to the node is as follows: (8) In equation (8), For the first item in the plan k pump truck r Each work location This is the sampling input node corresponding to this position. and These are the start and end times of the task period; within this period, data is written directly to the corresponding node. Write 0 at other times.

[0035] When the operating times of multiple pump trucks overlap at the same pumping input node, the total set flow rate of that node is: (9) The continuous time flow is calculated according to the model. t n Discretize to obtain the first i The sampling input node at the ... n The set flow rate at each calculation time: (10) Equation (9) is used to superimpose the flow rates of multiple pump trucks when parallel operation within the same grid is allowed; when parallel operation is not allowed, the same node at the same time corresponds to at most one operation period. Equation (10) only indicates that the node flow rate generated by the complete operation plan table is read according to the calculation time of the model, and no additional location judgment variables are constructed. Preferably, the rated pumping flow rate of each pump truck is 0.625 m³ / s.

[0036] Organize the pumping flow rates of all two-dimensional grids across all computational time points into a time series matrix: (11) In Equation (11), the matrix rows correspond to the pumping input nodes or their mapped two-dimensional grids, and the columns correspond to the calculation time. The matrix is ​​directly generated from the complete operation plan table of each pump truck and organized into time-series data readable by the hydrodynamic model according to the mapping relationship between grid number and node number. In specific implementation, a wide CSV table with time as the row and node number as the column can be used, or a long CSV table with three fields: calculation time, node number, and flow rate can be used, or other time-series data formats specified by the model can be used.

[0037] ⑤ Embedding two-dimensional hydrodynamic calculations in the pumping sequence

[0038] The two-dimensional model reads the pumping flow rate of the corresponding node at each computation time step. Let the theoretical pumping volume be... To set the product of the absolute value of the pumping flow rate and the time step, the available water volume in the current grid is: The actual effective extraction volume is: (12) The actual effective pumping flow rate is calculated from the actual effective pumping volume: (13) Equations (12) to (13) ensure that the actual pumping volume within a calculation time step does not exceed the water accumulation volume that the grid can provide, thereby avoiding negative water depth after pumping.

[0039] set up To be compatible with grid Adjacent sets of two-dimensional grids, For rainfall and surface runoff input, For adjacent grids j Inflow grid i Traffic, For grid i Flow to adjacent grids j Traffic, Given the exchange flow between the one-dimensional drainage system and the two-dimensional surface, the grid net water volume input rate excluding pump truck drainage is: (14) Substituting the actual effective pumping flow rate of the pump truck and the net water input rate of the grid into the water balance equation, the grid water depth is updated as follows: (15) because If the value is not greater than 0, the operation of the pump truck reduces the water depth of the target grid and affects the evolution of water accumulation in the surrounding area through subsequent water exchange between grids and one-dimensional-two-dimensional exchange.

[0040] ⑥ Single comprehensive water accumulation impact target

[0041] Let the water depth ranges for light, moderate, and severe flooding be [ ] h l ,h m ), [ h m ,h h )and[ h h (,+∞). When When located in the interval corresponding to level c, the level indicator function Select 1 if the value is 1, otherwise select 0. i The cumulative duration for each grid to be at water accumulation level c is: (16) Based on the weights for mild, moderate, and severe waterlogging, the [number] [unit] was [determined / classified]. i We weight the duration of each level in each grid to obtain: (17) The weighted water accumulation impact of each grid is summed according to the grid area to obtain the single comprehensive water accumulation impact target value: (18) In equation (18), the grid area A i The system reflects the extent of water accumulation; different water accumulation levels are weighted to reflect the impact of water depth; and the cumulative duration reflects the duration of water accumulation. For equal-area grids, A i These are common constants and can be omitted when ranking the schemes. Preferably, h l = 0.15 m, h m = 0.30 m, h h = 0.50 m, w l = 0.2 、w m = 0.3 、w h = 0.5 。

[0042] Compared to the baseline scenario without a pump truck, the overall improvement rate of flooding impact for Option X is: (19) Equation (19) is used to evaluate the degree of improvement of the optimized scheme relative to the scenario without pump truck, and does not constitute a second optimization objective.

[0043] ⑦ Overall optimization of the complete work plan

[0044] Based on the aforementioned rules for generating a complete work plan, generate randomly. N p A set of scheduling schemes that satisfy the constraints of candidate location, total job cycle, minimum job time, and mobility reachability constitutes the initial population: (20) Each scheduling scheme X contains a complete work plan for K pump trucks. The optimization algorithm considers the overall impact of water accumulation. J (X) is a single objective, solved within the set Ω of all feasible complete work plans: (twenty one) Each of the equations (20) All are scheduling schemes that include a complete work plan table for all pump trucks. Equation (21) represents the scheme that minimizes the overall impact of water accumulation from all feasible complete plans. The generation of the next work position is only used to construct a complete plan table that meets the time and movement constraints. The optimization algorithm compares and updates the entire complete work plan table, rather than optimizing the next position one by one during the operation.

[0045] ⑧ Travel time, reachability, and scheduling constraints

[0046] like Figure 4 As shown, the generation of the complete work schedule must satisfy constraints such as movement time, total work cycle, minimum work time, location accessibility, and work conflicts. When the two-dimensional grid uses row and column numbering, the Manhattan distance between adjacent work locations is: (twenty two) In equation (22), The side length of the two-dimensional orthogonal mesh. and The work location is respectively The corresponding grid row and column numbers.

[0047] Let the pump truck's moving speed be... v k The smallest discrete unit of travel time is The travel time is: (twenty three) No. k The operating time and travel time of each pump truck meet the total operation cycle constraint: (twenty four) The duration of the operation at any location satisfies: (25) When the remaining available time is At that time, the reachable set of the next job location is: (26) In equation (26), For the pump truck from its current location Move to candidate position The required discrete travel time is calculated according to equation (23).

[0048] In implementations that do not allow multiple pump trucks to occupy the same work grid at the same time, when two work plans correspond to the same grid, their work time intervals must not overlap, i.e., the following must be satisfied: (27) Preferably, the total work cycle T is 24 hours, and the minimum work time is... The time interval is 0.5 h, the pump truck's moving speed is 40 km / h, and the moving time is discrete in units of 5 min. If engineering conditions allow multiple pump trucks to be parked and operated in parallel, the time non-overlapping restriction of equation (27) can be cancelled, and the pumping flow rate can be superimposed according to equation (9).

[0049] ⑨ Single-objective optimization algorithm and termination condition

[0050] like Figure 6 As shown, a single-objective genetic algorithm based on an elite retention strategy in Geatpy is used to solve the problem. Each individual in the initial population consists of a complete operation plan table for K pump trucks. For any individual to be evaluated, it is first decoded into a complete operation plan table for each pump truck. Based on the plan table, a pumping flow time series is generated, and a one-dimensional to two-dimensional hydrological and hydrodynamic model is invoked. Subsequently, the comprehensive water accumulation impact target value is calculated based on the simulation results. J (X). Conditions such as candidate location, total job cycle, minimum job time, and mobility reachability are directly satisfied during the generation of the complete schedule. If an infeasible solution arises after crossover or mutation, it is repaired through the schedule to re-satisfy the preset scheduling conditions. No additional penalty function is set during objective function evaluation, and constraint violation is set to 0.

[0051] Comprehensive impact of water accumulation J (X) is set as the minimization objective, and Geatpy evaluates individual fitness based on the objective function value. The algorithm employs tournament selection, two-point crossover, and Breeder GA mutation, and forms the next generation population through parent-child merging and elite retention. Crossover and mutation are applied to the position sequence or job time parameters in the complete job schedule table; after position updates, illegal positions are corrected to the candidate job position set, and movement time and reachability are recalculated based on the updated positions; after time updates, time periods below the minimum job time are adjusted and the total job cycle is rebalanced. The corrected complete job schedule table is then re-evaluated using temporal embedding and hydrodynamic modeling.

[0052] The iteration terminates when the number of evolutionary generations reaches the preset maximum, or when the absolute value of the difference between the historical best objective value and the current generation's best objective value is less than the stagnation threshold and the stagnation count reaches the preset upper limit, and the historical best complete job plan is output. The maximum number of evolutionary generations, the stagnation threshold, and the upper limit of the stagnation count are set based on the number of candidate positions, the dimensions of the objective function, and available computing resources.

[0053] ⑩ Specific parameters and output results

[0054] In one specific embodiment, the two-dimensional surface model uses a 15 m × 15 m orthogonal grid, the number of mobile drainage pump trucks is 5, the rated pumping flow rate of a single pump truck is 0.625 m³ / s, the total operation cycle is 24 h, the minimum operation time is 0.5 h, the moving speed is 40 km / h, and the moving time discrete unit is 5 min.

[0055] In the initial population generation phase, each pump truck first randomly selects its first work location from the candidate work location set. Then, based on the remaining time, movement speed, and reachability, it generates subsequent work locations, work times, and movement times until a complete 24-hour work plan is formed. The complete work plans of all five pump trucks together constitute an optimized individual. When evaluating each individual, the pumping input nodes of different work grids switch between -0.625 m³ / s and 0 sequentially according to the arrival, work, and movement time periods in the plan, and are input into the two-dimensional model in a time-series data format readable by the hydrodynamic model; in specific implementations, CSV files can be used.

[0056] The output full-process scheduling plan includes pump truck number, initial work grid number, subsequent work grid numbers, arrival time, work start time, work end time, movement time, set pumping flow rate, and actual effective pumping flow rate; at the same time, it outputs a CSV pumping flow rate time series file corresponding to the complete work plan table for two-dimensional model recalculation and scheduling scheme verification.

[0057] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0058] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for optimizing the scheduling of mobile drainage pumping trucks based on pumping sequence embedding, characterized in that, include: Acquire basic data of the target area, and construct an urban waterlogging hydrological and hydrodynamic model based on the basic data, including a one-dimensional drainage network model and a two-dimensional surface runoff model. The historical actual water accumulation locations in the target area are summarized with the water accumulation locations simulated by the urban waterlogging hydrological and hydrodynamic model. A candidate operation location set is formed through preliminary manual screening. A mapping relationship is established between the two-dimensional grid numbers in the candidate operation location set and the one-dimensional model sampling input node numbers. The initial working position of each pump truck is randomly generated from the candidate working position set, and a complete work plan table containing subsequent working positions, working duration and moving time is generated based on the remaining working time, pump truck moving speed, minimum working time and location reachability constraints. Combine the complete work schedules of all pump trucks into a single scheduling scheme to be optimized. According to the complete work plan, the rated drainage flow for each work period is converted into a time-varying negative flow sequence for the corresponding drainage input node, and the time-varying negative flow sequence is input into the two-dimensional surface runoff model. At each calculation time step, the actual pumping volume is determined based on the available water volume in the current two-dimensional grid, the water depth in the two-dimensional grid is updated, and the target value of the single comprehensive water accumulation impact is calculated. A single-objective optimization algorithm is adopted, and the complete work plan table is used as the optimization individual for iterative updates. The output is the full-process scheduling plan of the mobile drainage pumping truck when the comprehensive water accumulation impact target value is minimized.

2. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The two-dimensional surface runoff model uses the node flow time series in the one-dimensional drainage network model as the water input interface; the two-dimensional grid number corresponds one-to-one with the pumping input node number, and the pumping input node is used to carry the pumping flow time series of the mobile drainage pump truck at the corresponding two-dimensional grid.

3. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping sequence embedding as described in claim 2, characterized in that, The one-dimensional drainage network model is constructed using SWMM. The pumping input node is implemented by adding manhole nodes corresponding to the candidate two-dimensional grid in the SWMM model. When the pump truck is dispatched to the target two-dimensional grid, the pumping flow time series of the corresponding manhole node is activated. The two-dimensional surface runoff model reads the pumping flow time series and applies the corresponding flow to the target two-dimensional grid.

4. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The candidate operation location set is formed by summarizing the historical actual water accumulation locations and the water accumulation locations simulated by the hydrological and hydrodynamic model in the target area, and then conducting a preliminary manual screening. The preliminary manual screening is used to eliminate locations that are obviously unsuitable as operation points for mobile drainage pumping vehicles, and the screening results are saved through a candidate grid marker file.

5. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The complete work plan table for each pump truck covering the preset total work cycle includes: multiple work grids arranged in an orderly manner, the work duration in each work grid, and the travel time between adjacent work grids; the first work location is randomly generated from the set of candidate work locations, and subsequent work locations are generated from candidate locations that meet the remaining time and travel accessibility constraints, and the work duration and travel time together cover the preset total work cycle.

6. The optimized scheduling method for mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The scheduling plan for all pump trucks is composed of the work plans of each pump truck. The arrival time of each work position is derived recursively from the arrival time, work time and travel time of the previous position. At the same time, the work end time of a certain work position is calculated. Each pump truck starts from the first randomly generated work position and generates subsequent work positions, work duration and travel time in sequence by combining the remaining work time, travel speed and reachable range, until a complete work plan table covering the total work cycle is formed. The complete work plan tables of all pump trucks together form an optimization unit, and the first work position, subsequent positions and related time parameters will be evaluated and updated as a whole within the same optimization unit.

7. The optimized scheduling method for mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The transformation of the time-varying negative flow sequence is specifically as follows: The two-dimensional grid corresponding to each pump truck in each working period is directly read from the complete work plan. During the pump truck's working period, the flow rate of the pumping input node corresponding to the two-dimensional grid is set to the negative rated pumping flow rate of the pump truck. During the pump truck's moving period, the flow rate of the relevant pumping input node is set to 0. When multiple pump trucks are allowed to operate simultaneously in the same grid, the rated pumping flow rates of each pump truck are superimposed.

8. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The specific calculation of the target value for the impact of a single comprehensive water accumulation is as follows: First, define the water depth ranges corresponding to light, moderate, and severe water accumulation. Then, use the level indicator function to determine the water accumulation level of each grid at each calculation time and calculate the cumulative duration of each grid corresponding to different water accumulation levels. Based on the set three levels of water accumulation weights, the cumulative duration of each level within a single grid is weighted and summed to obtain the weighted water accumulation impact of a single grid. Finally, the weighted water accumulation impact of all grids is multiplied by the corresponding grid area and summed to form a single comprehensive water accumulation impact target value that simultaneously reflects the water accumulation range, water accumulation depth, and water accumulation duration.

9. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The single-objective optimization algorithm is a genetic algorithm based on an elite retention strategy; Multiple scheduling schemes that satisfy the constraints of candidate location, total operation cycle, minimum operation time, and mobility reachability are randomly generated and form an initial population. Each scheduling scheme contains a complete operation plan table for K pump trucks. Subsequently, the optimization algorithm takes the overall impact of water accumulation as the single objective and finds the scheme with the minimum overall impact of water accumulation in the set of all feasible complete operation plans. The generation of the next operation location is only used to construct a complete plan table that satisfies the time and mobility constraints. The comparison and update of the optimization algorithm are carried out on the entire complete operation plan table, rather than optimizing the next operation location one by one during the operation.

10. The method for optimizing the scheduling of mobile drainage pumping vehicles based on pumping and drainage timing embedding as described in claim 1, characterized in that, The mobile drainage pump truck's full-process scheduling plan includes at least the pump truck number, initial work grid number, subsequent work grid numbers, arrival time, work start time, work end time, movement time, set drainage flow rate, actual effective drainage flow rate, and the corresponding drainage flow rate time series file.

Images