A port shore power berth construction multi-objective collaborative optimization planning method, device and medium

Abstract

The present application provides a kind of port shore power berth construction multi-objective collaborative optimization planning method, equipment and medium, the method includes: obtaining multi-source fusion data;With binary decision variable to represent whether berth is configured shore power facility;Four optimization objective functions of minimum resource consumption, carbon emission reduction benefit maximization, shore power utilization rate maximization and planning orientation adaptation degree maximization are constructed;Set resource total amount upper limit, berth type and decision variable constraint;Standardized processing is carried out to objective function;Multi-objective optimization model is constructed;Improved NSGA-III algorithm is used to solve Pareto optimal configuration solution set;Based on combination weight method, the configuration solution with the highest comprehensive evaluation value is selected as the final configuration strategy;The final strategy is subjected to sensitivity test and anti-disturbance ability verification.The present application realizes the multi-objective collaborative optimization of shore power construction economy, environment, efficiency and policy orientation, and improves the robustness and spatial balance of planning scheme under uncertain environment.

Description

Technical Field

[0001] This invention relates to the field of industrial internet technology, and is a typical application of industrial internet platform layer technology in the field of port green energy construction. In particular, it relates to a multi-objective collaborative optimization planning method, equipment and medium for port shore power berth construction. Background Technology

[0002] With the deepening of global "dual carbon" goals, ports, as international trade hubs and key carbon emission areas, have made the construction of shore power facilities a crucial measure to reduce ship berthing emissions and achieve green port transformation. Shore power technology, by supplying electricity from shore to moored ships, replaces fuel-fired power generation for ship auxiliary engines, significantly reducing emissions of sulfur oxides, nitrogen oxides, and carbon dioxide. However, the construction of shore power berths in ports involves significant investment in berth renovation, power grid expansion, equipment procurement and installation, resulting in high construction costs. Furthermore, it is influenced by multiple factors such as the availability of ship power receiving facilities, berth utilization rates, and policy guidance. How to scientifically select sites and rationally plan layouts within a limited budget to achieve synergistic optimization of economic, environmental, efficiency, and policy objectives is a pressing technical challenge in the field of port planning.

[0003] In existing technologies, optimization research on shore power systems largely focuses on the operational scheduling level. For example, some methods construct multi-objective models considering ship time costs, electricity costs, and shore power operation time costs, and employ evolutionary algorithms to solve for dynamic scheduling and berth allocation schemes for mobile shore power equipment, thereby improving the flexibility and economy of daily port operations. These studies typically assume that shore power facilities are already in place and focus on optimizing real-time operations with existing resources, falling under the category of short-term operational decision-making.

[0004] However, during the planning and construction phase of shore power facilities, decision-makers face entirely different technical challenges: how to strategically select berths for shore power facility construction to balance long-term investment benefits, environmental benefits, and policy requirements. Existing planning methods often focus on single-objective optimization, such as constructing linear programming models guided by cost minimization or emission reduction maximization, failing to adequately consider diverse decision-making needs such as shore power utilization efficiency and policy adaptability, making it difficult to implement planning schemes in actual operation. Some studies have attempted to introduce multi-objective optimization algorithms, but when faced with more than three optimization objectives, traditional algorithms suffer from uneven solution set distribution and insufficient convergence, making it difficult to provide a balanced Pareto front solution. Furthermore, existing technologies do not adequately consider the dynamic correlation between ship power receiving facilities and shore power berth construction. Planning models often assume that all berthed ships can use shore power, ignoring the limitations of the ships' own power receiving capabilities, resulting in significant discrepancies between carbon emission reduction benefits and shore power utilization rate estimates.

[0005] In terms of data support, traditional methods often rely on statistical data from a single source, lacking a systematic integration of basic berth information, ship berthing patterns, and policy and technical parameters. This data fragmentation is a significant problem, failing to provide accurate and traceable input for optimization models. In the decision-making stage, weight determination often employs either the analytic hierarchy process (AHP) or the entropy weight method. The former is highly subjective and easily influenced by expert preferences, while the latter relies entirely on data objectivity and struggles to reflect policy guidance; both have limitations. Furthermore, existing planning schemes generally lack sensitivity analysis and robustness verification for fluctuations in core parameters. When construction costs, berthing frequency, and load factors fluctuate in actual operation, the planning scheme may become ineffective.

[0006] In summary, there is an urgent need for a port shore power berth construction planning method that can integrate multi-source data, take into account multiple objectives, has strong algorithm adaptability, scientifically determined weights, and robust verification capabilities. This method would address the problems of single objectives, fragmented data, poor algorithm adaptability, and insufficient practicality of existing technologies, and provide reliable technical support for the scientific layout and sustainable development of port shore power facilities. Summary of the Invention

[0007] To address the aforementioned technical problems, the technical solution adopted by this invention is as follows:

[0008] According to a first aspect of the present invention, a multi-objective collaborative optimization planning method for the construction of shore power berths in ports is provided, the method comprising the following steps:

[0009] S100, acquire multi-source fusion data, which includes berth basic information, ship-berth berthing correlation data, planning guidance technical parameters, and sensitivity analysis parameters.

[0010] S200, based on the multi-source fusion data, binary decision variables are used to characterize whether each berth is equipped with shore power facilities.

[0011] S300 constructs at least several optimization objective functions, including minimizing resource consumption, maximizing carbon emission reduction benefits, maximizing shore power utilization, and maximizing planning guidance adaptability.

[0012] S400, based on the multi-source fusion data, the preset constraints of the multiple optimization objective functions, and the standardized objective function values, a multi-objective optimization model is constructed; wherein, the constraints are used to limit the feasible solution space of shore power facility configuration, and the standardization process is used to map objective values ​​of different dimensions to a unified scale.

[0013] S500 employs an improved NSGA-Ⅲ algorithm to solve the multi-objective optimization model within the feasible solution space defined by the constraints, thereby obtaining the Pareto optimal configuration solution set.

[0014] S600: Calculate the comprehensive evaluation value of each configuration solution in the Pareto optimal configuration solution set based on the combined weight method, and select the configuration solution with the highest comprehensive evaluation value as the final configuration strategy for shore power berth construction.

[0015] S700: Perform core parameter sensitivity testing and disturbance rejection capability verification on the final configuration strategy, evaluate whether the final configuration strategy still meets the constraints under parameter fluctuation conditions, and output a recommended configuration strategy that has been verified for feasibility.

[0016] According to a second aspect of the present invention, an electronic device is provided, including a processor and a memory; the processor executes the steps of the method described in the first aspect of the present invention by invoking a program or instructions stored in the memory.

[0017] According to a third aspect of the present invention, a computer-readable storage medium is provided that stores a program or instructions that cause a computer to perform the steps of the method described in the first aspect of the present invention.

[0018] The present invention has at least the following beneficial effects:

[0019] (1) Multi-objective collaborative optimization: By constructing four objective functions—minimizing resource consumption, maximizing carbon emission reduction benefits, maximizing shore power utilization, and maximizing planning guidance adaptability—economic efficiency, environmental benefits, operational efficiency, and policy requirements are incorporated into a unified optimization framework, which solves the problem that traditional methods have a single objective and are difficult to take into account the needs of multiple decision-making processes.

[0020] (2) Improved solution performance: The improved NSGA-Ⅲ algorithm is adopted, and an adaptive mutation mechanism and a dynamic reference point adjustment mechanism are introduced. The former dynamically adjusts the mutation probability according to the iteration process to balance the global search and local development capabilities, while the latter adjusts the reference point position and density in real time based on the population distribution state, which enhances the convergence and solution set distribution uniformity of the algorithm in high-dimensional discrete multi-objective optimization problems.

[0021] (3) Scientific scheme selection: The comprehensive evaluation value of each configuration solution is calculated by the entropy weight-fuzzy hierarchical analysis combined weight method. The fuzzy hierarchical analysis method integrates expert experience to determine subjective weights, while the entropy weight method determines objective weights based on data information entropy. The two are integrated according to a preset ratio so that the final selection scheme takes into account both expert judgment and data objectivity.

[0022] (4) Strong engineering robustness: The final configuration strategy is verified by basic verification, multi-condition joint verification and regional equilibrium stability verification. Parameter fluctuation scenarios are generated by Monte Carlo sampling and reduced to typical conditions by clustering. The scheme is tested to see if it still meets the constraints and maintains regional equilibrium under parameter fluctuations, so as to ensure that the planning scheme is feasible and robust in uncertain environments.

[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0025] Figure 1 The flowchart illustrates a multi-objective collaborative optimization configuration method for the construction of shore power berths in ports, as provided in an embodiment of the present invention. Detailed Implementation

[0026] 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.

[0027] Unless otherwise defined, 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 pertains. The terminology used herein in the description of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0028] It should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the steps as sequential processes, many of these steps can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the steps can be rearranged. A process can be terminated when its operation is complete, but it may also have additional steps not included in the figures. A process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0029] This invention provides a multi-objective collaborative optimization allocation method for the construction of shore power berths in ports. It aims to solve the problems of single optimization objectives, fragmented multi-source data, poor algorithm adaptability to high-dimensional objectives, and insufficient resilience of allocation strategies in traditional planning methods. This method achieves collaborative optimization of resource consumption, carbon emission reduction benefits, shore power utilization rate, and the fit of planning guidance during shore power construction. Figure 1 As shown, the method includes the following steps:

[0030] S100, acquire multi-source fusion data, which includes berth basic information, ship-berth berthing correlation data, planning guidance technical parameters, and sensitivity analysis parameters.

[0031] To ensure the comprehensiveness, traceability, and consistency of the input data, the multi-source fused data is organized and managed in the form of standardized database tables. The field definitions of each data table are as follows:

[0032] (1) Berth Basic Information Table, which describes the static attributes of each berth in the port, includes the following fields:

[0033] Berth identifier: A unique code for a berth, used to associate the same berth in different data tables;

[0034] Berth type: Berth categories classified by function, including but not limited to container berths, bulk cargo berths, general cargo berths, and passenger / roll-off berths;

[0035] Berth tonnage: The tonnage class of ships designed to berth at a berth, expressed in ten thousand tons;

[0036] Dedicated berth marking: Indicates whether the berth is a dedicated berth. A dedicated berth refers to a berth designed for a specific type of ship or cargo, such as a dedicated container berth, a dedicated ore berth, or a dedicated oil berth.

[0037] Shore power facility resource allocation consumption: The resource consumption required for configuring shore power facilities at this berth, including investments in equipment procurement, installation engineering, power grid transformation, etc., is measured in standardized resource units;

[0038] Total number of berthings per year: The total frequency of all types of vessels berthing at this berth within a historical year, expressed as times per year;

[0039] Port Area: The code of the area where the berth is located, used for spatial dimension analysis and balance optimization.

[0040] (2) Vessel-berth berthing information table, used to record dynamic data on vessel berthing behavior, includes the following fields:

[0041] Berth identifier: A unique code for a berth that is linked to the berth basic information table;

[0042] Ship type: Classified by ship purpose or tonnage, including but not limited to container ships, bulk carriers, tankers, and ro-ro passenger ships;

[0043] Annual berthing frequency of vessels equipped with power receiving facilities: The frequency of vessels equipped with power receiving facilities berthing at this berth during the year, expressed as times per year;

[0044] Average berthing time per trip: The average berthing time of this type of vessel at this berth, in hours;

[0045] Power of auxiliary engines for berthed vessels: The rated power of this type of auxiliary engine, expressed in kilowatts (kW).

[0046] (3) Planning-oriented technical parameter table, used to store parameters related to policy planning and technical specifications, including the following fields:

[0047] Upper limit of available resources for shore power construction: The upper limit of the total amount of resources that can be allocated for shore power construction in ports. The unit consumption of resources is consistent with the resource allocation and is used to constrain the overall construction scale.

[0048] Dedicated berth shore power coverage planning target: The planned proportion of dedicated berths with shore power configuration, expressed as a percentage (%);

[0049] Load factor: The ratio of the actual load of auxiliary machinery to the rated load during berthing, with a value ranging from 0 to 1, dimensionless;

[0050] CO2 emission factor: The carbon dioxide emission factor of auxiliary equipment fuel, expressed in kilograms of CO2 per kilogram of fuel. The specific value is determined according to the fuel type and with reference to national or industry standards.

[0051] (4) Sensitivity analysis parameter table, used to record the core parameters that need to be analyzed for volatility and their volatility range, including the following fields:

[0052] Parameter type: The category of parameters, including resource allocation consumption, number of times ships with power receiving facilities berth, load factor, etc.

[0053] Parameter name: The specific name of the parameter, corresponding to the field name in the aforementioned data table;

[0054] Baseline value: The initial value of the parameter;

[0055] Variation range: The percentage range of parameter fluctuation, such as ±10%, ±15%, etc.;

[0056] Values ​​after changes: Parameter values ​​calculated based on the magnitude of the changes, used to generate fluctuation scenarios.

[0057] To further improve the accuracy and timeliness of the data, the multi-source fusion data is constructed through cleaning, alignment, and correlation of heterogeneous data sources. Specifically, the raw data from different business systems (such as production scheduling systems, ship declaration systems, and historical statistical reports) are formatted, outliers are removed, and missing values ​​are filled. Various types of data are then correlated and aligned using key fields such as berth identifiers and ship types to form a fusion dataset containing historical berthing data and real-time dynamic data, providing complete and consistent data support for subsequent model optimization.

[0058] The selection criteria and technical functions of the aforementioned multi-source fusion data are explained below:

[0059] In the planning and construction of shore power berths in ports, the scientific validity and feasibility of the decision-making results highly depend on the completeness, accuracy, and relevance of the input data. Data from a single source or in isolation cannot support the complex decision-making needs of multi-objective collaborative optimization. Therefore, this invention constructs a multi-source fusion dataset from the following four dimensions, and the selection logic of data in each dimension and its technical role in the optimization model are as follows:

[0060] (1) Berth Basic Information: Berths are the basic units for shore power facility configuration, and their static attributes directly determine the feasibility of configuration decisions and resource consumption. Specifically, berth type and tonnage are used to determine whether the berth has the technical conditions for shore power configuration; dedicated berth identification is used to support coverage calculations in policy adaptability objectives; shore power facility resource consumption is the basic input for the objective function of minimizing resource consumption; the total number of berths per year reflects the berth's busyness, indirectly affecting the economic efficiency of shore power utilization; and the port area to which it belongs provides a basis for spatial equilibrium analysis. This dimension of data provides the physical carrier and constraint boundaries of decision variables for the optimization model.

[0061] (2) Ship-berth berthing correlation data: The emission reduction benefits and utilization rate of shore power depend on the actual power receiving capacity and energy demand of berthed ships. Among them, the annual number of berthings of ships with power receiving facilities is the core input for calculating carbon emission reduction benefits, directly determining the actual emission reduction potential of shore power; the average berthing time per berthing affects the duration of shore power use by ships, thus affecting the estimation of emission reduction; the auxiliary engine power of berthed ships is used to calculate the electrical load of ships during berthing. This dimension of data enables the optimization model to dynamically correlate ship-side characteristics with berth-side configuration, avoiding overestimation or underestimation of benefits due to neglecting the ship's power receiving capacity.

[0062] (3) Planning-Guided Technical Parameters: Port shore power construction must comply with industry standards and policy guidelines, while also being constrained by the total amount of available resources. Among these, the upper limit of the total available resources for shore power construction constitutes the core parameter constraining the total resource limit; the planning target for shore power coverage at dedicated berths serves as the quantitative basis for the planning-guided fit objective function; the load factor and CO2 emission factor are used to convert ship electricity demand into actual carbon emissions, ensuring the accuracy of emission reduction benefit calculations. This dimension of data transforms macroeconomic policy requirements and technical specifications into quantifiable constraints and targets for the optimization model.

[0063] (4) Sensitivity Analysis Parameters: In actual operation, core parameters such as resource allocation consumption, ship berthing frequency, and load factor are subject to uncertainty. If the planning scheme is too sensitive to parameter fluctuations, it may fail during execution. The sensitivity analysis parameter table records the core parameters that need to be analyzed for fluctuations and their fluctuation ranges, providing data support for subsequent verification of disturbance resistance capabilities. This dimension of data enables the optimization model to move from static optimality to dynamic reliability, improving the adaptability of the planning scheme in uncertain environments.

[0064] By integrating data from the above four dimensions, this invention constructs a complete data system covering the decision-making object (berth), demand side (ships), regulatory side (policy and technology), and uncertainty side (sensitivity analysis), providing a solid data foundation for the subsequent construction and solution of multi-objective optimization models.

[0065] S200, based on the multi-source fusion data, binary decision variables are used to characterize whether each berth is equipped with shore power facilities, and a mapping relationship between decision variables and berth entities is established.

[0066] Specifically, for each berth j (j=1,2,…,n) within the port, a binary decision variable x is defined. j ∈{0,1}, where:

[0067] x j =1 indicates that shore power facilities are configured at the j-th berth;

[0068] x j =0 indicates that no shore power facilities are installed at the j-th berth.

[0069] The decision variable x j Each berth corresponds one-to-one with the "berth identifier" field in the berth basic information table, forming the decision space of the optimization model. The decision variables for all berths constitute the decision vector x=[x1,x2,…,x…]. n This vector fully describes the spatial configuration of the port's shore power facilities.

[0070] The binary decision variables have the following technical functions:

[0071] As inputs to the objective functions: decision variables are directly substituted into the four objective functions of minimizing resource consumption, maximizing carbon emission reduction benefits, maximizing shore power utilization, and maximizing planning guidance adaptability, to determine the contribution of each berth to each optimization objective;

[0072] As a carrier of constraints: decision variables are simultaneously substituted into constraints such as the upper limit of total resources, berth type constraints, and spatial distribution constraints to determine whether candidate configuration schemes meet the feasibility requirements.

[0073] As the search object of the optimization algorithm: the improved NSGA-Ⅲ algorithm uses the decision vector as the search object, seeks optimization in the feasible solution space, and generates the Pareto optimal configuration solution set.

[0074] By defining the binary decision variables as described above, this invention transforms the planning problem of port shore power berth construction into a combinatorial optimization problem, laying the foundation for the subsequent construction and solution of a multi-objective optimization model.

[0075] S300 constructs at least several optimization objective functions, including minimizing resource consumption, maximizing carbon emission reduction benefits, maximizing shore power utilization, and maximizing planning guidance adaptability.

[0076] To balance the overall benefits and spatially balanced development of port shore power construction, and to avoid problems such as excessive concentration of resources, large differences in utilization rates, and uneven achievement of policy objectives, this invention introduces a regional equilibrium penalty term into the original objective function, constructing an improved multi-objective function system. The specific construction process is as follows:

[0077] (I) Construction of the original objective function

[0078] Based on the multi-source fusion data, four original objective functions were first constructed without considering regional equilibrium. The quantification logic of each objective function strictly conforms to the actual operation scenario of the port, ensuring the feasibility of the optimization results.

[0079] (1) Objective function for minimizing resource consumption

[0080] .

[0081] Where, x j Let C be the binary decision variable for the j-th berth. j The resource allocation consumption for shore power facilities at the j-th berth is denoted by , and the unit is consistent with the upper limit of the total available resources; n is the total number of port berths; and j takes values ​​from 1 to n.

[0082] (2) Objective function for maximizing carbon emission reduction benefits

[0083] .

[0084] Where m is the number of ship types, SP ij T represents the number of times a vessel of class i equipped with power receiving facilities berths at berth j, expressed as times per year; ij P represents the average berthing time per trip for vessels of class i, in hours; i The rated power of the auxiliary engine for Class i ships, in kilowatts; ρ i τ is the load factor, dimensionless, and ranges from 0 to 1. i Fuel consumption rate, expressed in kilograms per kilowatt-hour; EF i The CO2 emission factor for auxiliary equipment fuel is expressed in kilograms of CO2 per kilogram of fuel; the value of i ranges from 1 to m.

[0085] The quantitative logic of the objective function for maximizing carbon emission reduction benefits lies in the fact that the calculation of carbon emission reductions must be linked to the ship's power receiving facilities. Actual emission reductions are only counted when a ship equipped with power receiving facilities berths at berths already equipped with shore power facilities. Therefore, the formula uses SP. ij The number of berths should be calculated based on (the number of berths of ships equipped with power receiving facilities) rather than the number of berths of all ships. This ensures the accuracy of the emission reduction benefit estimation and avoids overestimation of benefits due to neglecting the ships' own power receiving capabilities.

[0086] (3) Objective function for maximizing shore power utilization

[0087] .

[0088] Among them, Q ij This represents the total number of times a vessel of type i berths at the j-th berth, expressed as times per year.

[0089] The quantitative logic of the objective function for maximizing shore power utilization lies in the following: the numerator is the total number of berths of ships equipped with power receiving facilities, and the denominator is the total number of berths of all types of ships. By calculating the average utilization rate of each berth, the focus is on improving the overall shore power utilization rate of the port. This quantitative method can truly reflect the actual utilization level of shore power facilities and avoid diluting the utilization rate indicator by including ships without power receiving facilities.

[0090] (4) Planning-oriented adaptability maximization objective function

[0091] .

[0092] Where S is the set of dedicated berths, and |S| is the total number of dedicated berths.

[0093] The quantitative logic of the objective function for maximizing the planning guidance fit further includes: to reflect the relative importance of different dedicated berths under policy guidance, a priority weight w for dedicated berths is introduced. jThe weights are determined using the group decision analytic hierarchy process (AHP), specifically including: integrating the judgment matrices of at least five industry experts; performing a consistency check on the judgment matrices; and then defuzzifying and normalizing the judgment matrices that pass the consistency check to obtain the priority weights for each dedicated berth or dedicated berth type, satisfying w j ∈[0,1] and the sum is equal to 1. This priority weight is used for the weighted coverage calculation in the subsequent area equalization penalty term.

[0094] (II) Improving the objective function by introducing a regional equilibrium penalty term

[0095] To achieve coordinated optimization of shore power construction in the spatial dimension and prevent excessive concentration of resource allocation and benefit distribution in a few areas, this invention introduces a regional equilibrium penalty term into the original objective function to form an improved objective function.

[0096] (1) Port area division

[0097] Based on the port's geographical layout, administrative management, or functional zoning, the port is divided into K port areas, each port area r containing a number of berths, denoted as set G. r r=1,2,…,K; The basis for the division of port areas includes, but is not limited to: wharf operation area, berth tonnage distribution, cargo type, environmentally sensitive areas, etc.

[0098] (2) Determination of the regional balance coefficient

[0099] Using the analytic hierarchy process (AHP) or expert scoring method, and considering factors such as historical berthing frequency, environmental sensitivity, and planning guidance requirements for each area, the equilibrium coefficient λr for each area is determined, satisfying: λ k ∈[0,1] and ∑ r=1 K (λ) r =1.

[0100] Equilibrium coefficient λ r This reflects the relative importance of the r-th area in the overall plan and is used to balance the development needs of each area during the optimization process.

[0101] (3) Improved objective function

[0102] Based on the original objective function, regional equilibrium penalty terms are introduced for the three objectives of minimizing resource consumption, maximizing shore power utilization, and maximizing planning guidance fit, while the objective of maximizing carbon emission reduction benefits remains unchanged. Specific improvements are as follows: Resource consumption minimization objective function (improved):

[0103] .

[0104] Among them, AvgC rThe expected resource allocation consumption for the r-th region can be determined based on historical investment data or planned budget allocation for that region; μ1 is the resource consumption equilibrium penalty coefficient, used to control the weight of the cost deviation term in the objective function. This formula minimizes the total resource allocation consumption while forcing the actual consumption of each region to approach the expected value through the penalty term, thus achieving spatial equilibrium of resource input.

[0105] The objective function for maximizing shore power utilization (improved):

[0106] .

[0107] in, , where is the average shore power utilization rate of the r-th area, AvgU is the average shore power utilization rate of all areas, and μ2 is the utilization rate equalization penalty coefficient; this formula improves the overall shore power utilization rate while suppressing the utilization rate gap between areas through the penalty term, thus promoting the balanced utilization of shore power resources.

[0108] The improved objective function for maximizing the fit of the planning direction is as follows:

[0109] .

[0110] in, Let P be the coverage rate of dedicated berths in the r-th area. If there are no dedicated berths in this area, then P is... r =0, AvgP is the average dedicated berth coverage rate of all areas, and μ3 is the coverage rate equilibrium penalty coefficient; this formula maximizes the overall dedicated berth coverage rate while promoting the balanced achievement of policy objectives in each area through the penalty term.

[0111] The objective function for maximizing carbon emission reduction benefits remains unchanged because its benefits are inherently related to the distribution of ship berthing, and there is no need to introduce additional regional equilibrium penalties.

[0112] (III) Determination and Adjustment of the Balanced Penalty Coefficient

[0113] The equilibrium penalty coefficients μ1, μ2, and μ3 are used to control the relative weights of the region equilibrium penalty terms in the objective function. Larger coefficient values ​​indicate a higher requirement for the equilibrium of each region in the optimization result; smaller coefficient values ​​indicate a result closer to the optimal solution of the original objective function.

[0114] The methods for determining the equilibrium penalty coefficient include:

[0115] Dynamic adjustment based on decision preferences: Decision-makers can adjust the coefficient values ​​in different optimization rounds according to the port development strategy, observe the changes in equilibrium, and obtain a satisfactory configuration scheme.

[0116] Determine a reasonable range through sensitivity analysis: Conduct grid search or experimental design within the range of coefficient values ​​to analyze the impact of different coefficient values ​​on the Pareto solution set distribution and regional equilibrium index, and select a coefficient range that can maintain overall benefits and ensure spatial equilibrium.

[0117] In the improved NSGA-Ⅲ algorithm, the adjusted equilibrium penalty coefficient is substituted into the improved objective function. Under the premise of satisfying the constraints, the decision vector x is optimized to generate a Pareto optimal configuration solution set that takes into account both overall benefits and spatial equilibrium.

[0118] The present invention sets the above four optimization objective functions based on a systematic analysis of the planning problem for the construction of shore power berths in ports. It aims to solve the shortcomings of traditional methods, such as single objective and insufficient coordination. The specific reasons are as follows:

[0119] (1) Reasons for setting the goal of minimizing resource consumption

[0120] The construction of shore power facilities involves multiple investments, including equipment procurement, installation, and power grid upgrades. Given limited port resources, construction cost is the primary constraint in decision-making. Considering only environmental benefits or utilization rates while neglecting resource input could lead to planning schemes exceeding the port's carrying capacity and becoming unfeasible. Therefore, minimizing resource consumption is one of the objectives to ensure the economic feasibility of the optimized results.

[0121] (2) Reasons for setting the target of maximizing carbon emission reduction benefits

[0122] Under the "dual carbon" target, ports are key carbon emission areas, and the main driving force for shore power construction is to reduce carbon emissions during ship berthing. Focusing solely on construction costs or utilization rates while ignoring actual emission reduction effects may result in shore power facilities failing to effectively reduce carbon emissions after completion, deviating from the policy's original intent. Therefore, maximizing carbon emission reduction benefits is the core environmental objective, guiding resources towards berths with the greatest emission reduction potential. The inclusion of ship power receiving facility coverage in the formula is because only ships equipped with power receiving facilities can actually use shore power to generate emission reductions. This setting avoids overestimating emission reduction benefits due to neglecting ship-side capabilities.

[0123] (3) Reasons for setting the target of maximizing shore power utilization

[0124] The emission reduction benefits of shore power facilities ultimately need to be realized through actual use. If the utilization rate of berths is low after construction, it not only wastes resources but also fails to achieve environmental goals. Therefore, maximizing shore power utilization is used as an efficiency objective, guiding the priority construction of berths with high frequency and long berthing time for vessels equipped with power receiving facilities. The numerator in the formula is the number of berths by vessels equipped with power receiving facilities, not the total number of berths, because only vessels equipped with power receiving facilities are likely to use shore power; the denominator is the average of the total number of berths by all vessels, which accurately reflects the actual utilization level of shore power facilities.

[0125] (4) Reasons for setting the goal of maximizing the adaptability of the planning orientation

[0126] The construction of shore power facilities at ports must comply with industry policies and planning requirements, especially for dedicated berths (such as container and roll-on / roll-off berths), which are often the priority targets of policy coverage. If the planning scheme deviates from policy guidance, it may lead to difficulties in subsequent approvals or insufficient policy support. Therefore, maximizing the alignment with planning guidance is taken as the policy objective, guiding the priority construction of shore power facilities at dedicated berths. The formula introduces a priority weight for dedicated berths (determined through AHP group decision-making) because different dedicated berths have varying importance in policy planning; weighted processing ensures that the planning results are more closely aligned with actual policy requirements.

[0127] (5) The necessity of synergistic coordination of four objectives

[0128] The four objectives mentioned above characterize the planning and construction of shore power from four dimensions: economic feasibility, environmental benefits, operational efficiency, and policy compliance. Optimizing a single objective cannot simultaneously consider other dimensions; for example:

[0129] Considering only economic feasibility may result in low utilization rate after construction;

[0130] Environmental benefits alone may exceed resource carrying capacity;

[0131] Considering only operational efficiency may overlook key policy areas;

[0132] Considering only policy compliance may lead to a waste of resources.

[0133] By constructing a four-objective collaborative optimization model, this invention can balance economic, environmental, efficiency, and policy requirements under resource constraints and output a comprehensive optimal configuration scheme.

[0134] S400, based on the multi-source fusion data, the multiple optimization objective functions, preset constraints, and standardized objective function values, a multi-objective optimization model is constructed, specifically including:

[0135] S410, Set constraints

[0136] In this invention, the preset constraints include a maximum resource limit, berth type constraints, and decision variable constraints, which are used to limit the feasible solution space for shore power facility configuration.

[0137] To ensure the feasibility of the planning scheme in engineering practice, this invention sets the following four types of constraints, which limit the feasible solution space from the dimensions of total resources, berth attributes, spatial distribution, and decision variables:

[0138] (1) Maximum resource limit constraint

[0139] There is an upper limit to the total amount of resources that can be allocated for the construction of shore power facilities in ports. The sum of the resource consumption for all shore power facilities configured at all berths must not exceed this upper limit. The mathematical expression for this is: .

[0140] Among them, B max This sets a maximum limit on the total amount of resources available for shore power construction; this constraint ensures that the planning scheme does not exceed the port's resource carrying capacity.

[0141] (2) Berth type constraints

[0142] According to port development plans and shore power promotion requirements, different types of berths must meet specific configuration requirements. These berth type constraints include:

[0143] Configuration requirements for dedicated berths and general-purpose berths: In accordance with the shore power promotion policy, it may be stipulated that dedicated berths must be equipped with shore power facilities, general-purpose berths are prohibited from being equipped with shore power facilities, or mandatory configuration requirements may be set for specific types of berths.

[0144] Berth tonnage suitability constraint: The design tonnage of berths equipped with shore power facilities must match the tonnage of the vessels planned to berth there. In other words, a berth can only provide shore power services to a certain type of vessel if its tonnage is not lower than the required tonnage. This constraint is implicit in the calculation of carbon reduction benefits and shore power utilization rates, because only berths with tonnage matching can actually serve the corresponding vessels. This constraint is implemented in the optimization model through conditional judgments during the objective function calculation; that is, when the berth tonnage and the vessel tonnage do not match, the relevant benefit terms are not included in the objective function value.

[0145] (3) Spatial distribution constraints

[0146] To avoid excessive concentration of shore power facilities in a few areas and to ensure balanced development across port areas, each port area is required to have at least one shore power berth, expressed mathematically as: ∑ j∈Gr x j ≥1. This constraint mandates that each area has shore power service capabilities.

[0147] (4) Decision variable constraints: x j ∈{0,1}, meaning it can only take the values ​​0 or 1.

[0148] This constraint defines the solution space property of the optimization problem, namely, whether or not to configure shore power facilities for each berth is a discrete decision.

[0149] The above four types of constraints together constitute the feasible solution space for shore power facility configuration. Any decision vector x that satisfies all constraints is a feasible solution, providing clear boundary conditions for subsequent multi-objective optimization solutions.

[0150] S420, standardize the multiple optimization objective functions to map the objective values ​​of different dimensions to a unified scale.

[0151] Since the four objective functions have different dimensions and orders of magnitude—resource consumption measured in resource units, carbon emission reduction benefits measured in tons of carbon dioxide, and shore power utilization rate and planning guidance fit expressed as percentages—directly substituting them into a multi-objective optimization model would lead to the larger-dimensional objective dominating the solution process, affecting the distribution balance of the Pareto front and the diversity of the solution set. Therefore, this invention standardizes the multiple objective functions, with the specific steps as follows:

[0152] (1) Determine the range of values ​​for each objective function.

[0153] For each optimization objective function f p (p=1,2,3,4), firstly, through single-objective optimization or historical data statistics, determine its theoretical minimum value f in the feasible solution space. min p and maximum value f max p :

[0154] f min 1. f max 1: The minimum and maximum possible values ​​of the resource consumption minimization objective under all constraints;

[0155] f min 2. f max 2: The minimum and maximum possible values ​​of the carbon emission reduction benefit maximization target;

[0156] f min 3, f max 3: The minimum and maximum possible values ​​of the shore power utilization maximization target;

[0157] f min 4. f max 4: The minimum and maximum possible values ​​of the goal of maximizing the adaptability of the planning orientation.

[0158] (2) Standardization is performed using the range transformation method.

[0159] The range transformation method is used to map the values ​​of each objective function to the [0,1] interval, ensuring that all objectives participate in optimization at the same scale. The standardized formula is as follows:

[0160] For the minimization objective (resource consumption f1):

[0161] f norm 1 = (f1 - f) min 1) / (f) max 1-f min 1);

[0162] For maximizing the objectives (carbon emission reduction benefit f2, shore power utilization rate f3, planning guidance fit f4):

[0163] f norm q =(f max q -f q ) / (f max q -f min q ); q=2,3,4

[0164] After the above transformation, all standardized objective function values ​​belong to [0,1], and the smaller the value, the better (i.e., unified into a minimization form), which facilitates the unified processing of subsequent multi-objective optimization algorithms.

[0165] (3) Handling extreme values ​​and abnormal situations

[0166] During the standardization process, if the maximum and minimum values ​​of a certain objective function are equal (i.e., the objective is a constant in the feasible solution space), its standardized value is set to 0 (for the minimization objective) or 1 (for the maximization objective), or the objective is removed from the optimization model.

[0167] (4) Forming a standardized target vector

[0168] The four standardized objective function values ​​are used to construct a standardized objective vector: F norm =[f norm 1, f norm 2, f norm 3, f norm 4).

[0169] This vector serves as input to the multi-objective optimization model and participates in the solution process of the improved NSGA-Ⅲ algorithm. Standardization ensures that each objective has equal dimensionality during the optimization process, laying the foundation for generating a uniformly distributed and well-converged Pareto optimal solution set.

[0170] S430, construct a multi-objective optimization model.

[0171] After completing the acquisition of multi-source fusion data, definition of decision variables, construction of the optimization objective function (including the regional equilibrium penalty term), setting of constraints, and standardization of the objective function, this invention further integrates the above elements into a complete multi-objective optimization model. This model uses the decision vector x as the optimization variable, the standardized multiple objective functions as the optimization guide, and the constraints as the feasible region boundary, forming a standard multi-objective optimization problem statement.

[0172] (a) Model Input

[0173] The input to the multi-objective optimization model includes the following three parts:

[0174] Multi-source fusion data: berth basic information, ship-berth berthing correlation data, planning guidance technical parameters and sensitivity analysis parameters from step S100, providing basic data support for objective function calculation and constraint condition verification;

[0175] Standardized objective function: The standardized objective function f from step S420 norm 1, f norm 2, f norm 3, f norm 4. After unifying the units of measurement, a comparable multi-objective system is formed;

[0176] Constraints: The upper limit of total resources, berth type, spatial distribution and decision variable constraints from step S410 together define the feasible solution space.

[0177] (II) Model Composition

[0178] Based on the above inputs, the constructed multi-objective optimization model consists of the following three parts:

[0179] (1) Decision variables

[0180] The decision variable is a binary vector x. This vector fully describes the spatial configuration of the port's shore power facilities.

[0181] (2) Objective function

[0182] The optimization objective is to minimize the standardized objective function vector:

[0183] minF(x)=[f norm 1(x), f norm 2(x), f norm 3(x), f norm 4(x)]

[0184] in:

[0185] f norm 1(x) is the standardized resource consumption minimization objective, which is obtained by range transformation from the corresponding original objective;

[0186] f norm 2(x) is the standardized carbon emission reduction benefit maximization target, which is obtained by range transformation from the corresponding original target;

[0187] f norm 3(x) is the standardized shore power utilization maximization target, which is obtained by range transformation from the corresponding original target (including the area balance penalty term);

[0188] f norm 4(x) is the standardized planning orientation fit maximization target, which is obtained by range transformation from the corresponding original target (including the area balance penalty term).

[0189] This vector form indicates that the optimization objective is no longer a single scalar, but a multi-dimensional vector composed of four conflicting objectives. There is no single solution that simultaneously optimizes all objectives, and a Pareto optimal solution set needs to be sought through multi-objective optimization methods.

[0190] (3) Constraints

[0191] The feasible solution space is defined by the aforementioned four types of constraints. Any decision vector that satisfies all of the above constraints is a feasible solution, and all feasible solutions constitute the feasible solution space Ω.

[0192] In this invention, the multi-objective optimization model belongs to the high-dimensional discrete multi-objective combinatorial optimization problem, and has the following characteristics:

[0193] Discrete decision variables: Binary variables lead to a discrete solution space, making it impossible to directly use continuous optimization methods;

[0194] Nonlinearity of the objective function: The introduction of the area equilibrium penalty term makes the objective function exhibit nonlinear characteristics;

[0195] The constraints are diverse, covering multiple dimensions such as resources, type, and space, forming a complex feasible domain;

[0196] Conflicts between objectives: The four objectives are mutually constrained, and there is no absolutely optimal solution.

[0197] (III) Expected Model Output

[0198] The expected output of this multi-objective optimization model is the Pareto optimal solution set, which is all feasible solutions x that satisfy the following conditions. * The set of: There is no other feasible solution x such that F(x) is not inferior to F(x) on all objectives. * And it is strictly superior to F(x) on at least one objective. *This solution set will serve as the solution objective for subsequent steps S500 and the basis for scheme selection in S600.

[0199] Through the above model construction, this invention transforms the port shore power berth construction planning problem into a standard multi-objective optimization problem, laying a mathematical model foundation for subsequent solution using the improved NSGA-Ⅲ algorithm.

[0200] S500 employs an improved NSGA-Ⅲ algorithm to solve the multi-objective optimization model within the feasible solution space defined by the constraints, thereby obtaining the Pareto optimal configuration solution set.

[0201] This invention is an improvement upon the NSGA-III algorithm framework. NSGA-III is a reference-point-based non-dominated sorting genetic algorithm that guides the population towards a uniform distribution towards the Pareto front through preset reference points. Its core steps include: population initialization, non-dominated sorting, reference point association, environmental selection, and genetic operations. For a detailed explanation of the working principle of NSGA-III, please refer to existing literature; it will not be elaborated upon here.

[0202] This invention introduces a dual adaptive mechanism on top of the aforementioned algorithm—an adaptive mutation mechanism at the genetic operation level and a dynamic reference point adaptive adjustment mechanism at the environment selection level—to improve the convergence speed and solution set uniformity of the algorithm in high-dimensional discrete multi-objective optimization problems. The two mechanisms operate at different stages of the algorithm, synergistically optimizing the solution performance.

[0203] In the NSGA-III algorithm, each individual corresponds to a feasible solution (i.e., a shore power facility configuration scheme) in the decision space, and its quality is evaluated by the objective function value vector in the objective space. A reference point is set in the objective space to guide the population to distribute uniformly towards the Pareto front.

[0204] (I) Adaptive mutation mechanism

[0205] To balance the algorithm's global search capability with its local exploitation capability and prevent premature convergence, this invention introduces an adaptive mutation mechanism in the genetic operation phase. This mechanism dynamically adjusts the mutation probability based on the iteration progress, causing the mutation probability to decrease as the number of iterations increases. The specific implementation is as follows:

[0206] Initial stage: Set a higher mutation probability (e.g., 0.03~0.05) to enhance population diversity, expand the search range, and avoid the algorithm getting stuck in local optima;

[0207] Mid-stage: As iterations proceed, the mutation probability gradually decreases, causing the algorithm to gradually shift from global search to local refined search;

[0208] Final stage: The mutation probability decays to a low level (e.g., 0.01~0.02), with a focus on preserving the genetic structure of superior individuals to promote algorithm convergence.

[0209] Meanwhile, this invention uses the crowding distance method to maintain the uniformity of the Pareto solution set distribution, and stores the non-dominated solutions of each generation in an external archive, and performs pruning operations on the archive periodically to remove redundant solutions, ensuring that the final output Pareto front has good distribution characteristics.

[0210] (ii) Dynamic reference point adaptive adjustment mechanism

[0211] The standard NSGA-III algorithm uses pre-defined fixed reference points, which are prone to reference point failure or uneven solution set distribution when facing complex or irregular Pareto fronts. To address this, this invention introduces a dynamic adaptive reference point adjustment mechanism, which adjusts the position and density of reference points in real time based on the distribution of each generation of the population. The specific steps are as follows:

[0212] Population distribution status perception: Calculate the distribution boundary of the current population in the target space (minimum and maximum values ​​of each target) and the degree of aggregation (through density estimation or neighborhood distance measurement) to identify sparse and dense regions of the solution set.

[0213] The reference point coverage is adaptive: based on the actual distribution boundary of the current population in the target space, the coverage of the reference point set is dynamically adjusted. Specifically, the hyperrectangle formed by the minimum and maximum values ​​of the current population in each target dimension is used as the effective search region, ensuring that reference points are always distributed within this region. As the population evolves, when the population distribution range expands, the reference point coverage expands accordingly; when the population converges to a smaller region, the reference point coverage shrinks synchronously, ensuring that reference points are always focused on the effective region of the current Pareto front, avoiding the distribution of reference points in ineffective regions and reducing selection efficiency.

[0214] Reference point density adaptive adjustment: Based on the local clustering degree of the population in the target space, the distribution density of reference points is dynamically adjusted to balance the exploratory nature and uniformity of the solution set. The specific implementation is as follows:

[0215] Sparse region identification: The number of individuals associated with each reference point is counted, and the average number of associated individuals for all reference points is calculated. If the number of associated individuals for a reference point is lower than a preset proportion (e.g., one-third of the average), the region where the reference point is located is determined to be a sparse region, indicating that this local region of the current Pareto front has not been fully explored.

[0216] Dense region identification: Calculate the Euclidean distance between reference points. If the distance between two reference points is less than a preset threshold and the number of associated individuals of both is higher than the average number of associated individuals, then the region is determined to be a dense region. This indicates that multiple reference points are excessively concentrated in the same local area, which may lead to an imbalance in selection pressure.

[0217] Adaptive adjustment operation: In sparse regions, new reference points are generated through interpolation to guide the algorithm to search for areas that have not been fully explored, thereby improving the coverage of the Pareto front by the solution set; in dense regions, neighboring reference points are merged to avoid redundancy caused by too many reference points and to maintain the uniformity of the solution set distribution.

[0218] To determine whether reference points are too dense, this invention employs a dynamic threshold based on the average spacing between reference points. First, the pairwise Euclidean distances between all reference points in the standardized target space are calculated. For each reference point, its distance to its nearest neighbor is taken; the mean of the nearest neighbor distances for all reference points is then calculated. If the distance between two reference points is less than γ times the mean nearest neighbor distance, and the number of associated individuals for both is greater than the average number of associated individuals, then the region is determined to be a dense region, and these two reference points are merged into one reference point, for example, by taking the geometric center of the two reference points or by weighting the average based on the number of associated individuals. After merging, the distances and means need to be recalculated until there are no dense point pairs, avoiding newly generated dense regions after each merge. Here, γ is the density coefficient, ranging from 0.4 to 0.6. This dynamic threshold mechanism enables density recognition to adapt to the size and distribution of the reference point set, avoiding the problem of fixed thresholds failing at different iteration stages.

[0219] This mechanism enables the distribution of reference points to evolve dynamically with the population, effectively overcoming the problem of solution set offset or uneven distribution caused by improper setting of fixed reference points, thereby improving the convergence performance and solution set quality of multi-objective optimization solutions.

[0220] (III) Algorithm Solution Process

[0221] Based on the aforementioned dual adaptive mechanism, the improved NSGA-Ⅲ algorithm solves the multi-objective optimization model within the feasible solution space. The specific process is as follows:

[0222] Step 1, Initialize the population: Randomly generate an initial population containing N individuals, each individual corresponding to a decision vector x, representing a shore power facility configuration scheme.

[0223] Step 2, Fitness Evaluation: For each individual in the population, calculate its standardized objective function value f based on multi-source fusion data. norm 1, f norm 2, f norm 3, f norm 4. Check whether all constraints are met. Individuals that do not meet the constraints are penalized or eliminated.

[0224] Step 3, Non-dominated sorting: Perform a fast non-dominated sort on the population to divide individuals into different non-dominated levels (Pareto fronts).

[0225] Step 4, Reference point association: Based on the current distribution status of the population, a set of reference points is generated or adjusted through a dynamic reference point adaptive adjustment mechanism, and each individual is associated with the nearest reference point.

[0226] Step 5, Environmental Selection: A reference-point-based selection strategy is adopted to select N individuals from the merged set of the current population and the offspring population to form a new generation population. During the selection process, individuals with high non-dominant levels and good association with the reference point are given priority.

[0227] Step 6, Genetic Operations: Perform selection, crossover, and adaptive mutation operations on the selected individuals to generate the offspring population. The mutation probability is dynamically adjusted based on the current iteration number.

[0228] Step 7, External Archive Update: Store the non-dominated solutions in the current population into the external archive, and prune the archive using the crowding distance method to ensure that the archive size is controllable and the solution set is evenly distributed.

[0229] Step 8, Termination Judgment: Check whether the preset maximum number of iterations or convergence criteria have been reached (e.g., no significant improvement in Pareto fronts for multiple consecutive generations). If not met, return to Step 2; if met, output the non-dominated solution set in the external archive as the Pareto optimal configuration solution set.

[0230] (iv) Solution Results

[0231] The improved NSGA-Ⅲ algorithm described above yields a Pareto optimal configuration set, representing all non-dominant shore power facility configuration schemes. Each scheme in this set achieves a balance among four objectives: resource consumption, carbon reduction benefits, shore power utilization, and planning orientation fit. No single scheme outperforms the others in all objectives. This solution set serves as input to step S600, further refining the overall optimal final configuration strategy.

[0232] S600: Calculate the comprehensive evaluation value of each configuration solution in the Pareto optimal configuration solution set based on the combined weight method, and select the configuration solution with the highest comprehensive evaluation value as the final configuration strategy for shore power berth construction.

[0233] To select a final configuration scheme from the Pareto optimal solution set that balances expert experience and data objectivity, this invention employs the entropy-weighted fuzzy hierarchical analysis method to calculate the combined weights of each optimization objective function, and calculates the comprehensive evaluation value of each configuration solution based on the weighted summation method. The specific steps are as follows:

[0234] (1) Subjective weight calculation: Fuzzy analytic hierarchy process

[0235] The subjective weights of each optimization objective function are determined using fuzzy hierarchical analysis to reflect the decision-making experts' preferences for port development strategies, policy guidance, and practical operational experience.

[0236] Constructing a fuzzy judgment matrix: Several port planning experts were invited to conduct pairwise comparisons of the importance of four optimization objectives (resource consumption, carbon emission reduction benefits, shore power utilization rate, and planning guidance fit). The fuzzy judgments of the experts were quantified using triangular fuzzy numbers (L, M, U), where L is the lower limit, M is the median, and U is the upper limit.

[0237] Consistency check: Perform a consistency check on the fuzzy judgment matrix of each expert, and remove the judgment matrices that fail the check to ensure the logical consistency of the expert evaluation;

[0238] Defuzzification and Normalization: The fuzzy judgment matrix that has passed the consistency test is defuzzified (e.g., using the centroid method or mean method) to obtain the clear weights of each objective. Then, normalization is performed to obtain the subjective weight vector w of each optimization objective. sub =[w1 sub w2 sub w3 sub w4 sub And the sum of all subjective weights equals 1.

[0239] (2) Objective weight calculation: Entropy weight method

[0240] The objective weights of each optimization objective function are determined using the entropy weight method. Based on the distribution characteristics of the objective values ​​of each scheme in the Pareto solution set, the information content of the data itself is reflected.

[0241] Constructing the evaluation matrix: Using the Z best-case scenarios in the Pareto optimal solution set as samples, each solution corresponds to four standardized objective function values. Let the standardized value of the v-th best-case scenario under the p-th optimization objective be a. vp =f norm p (x v (p=1,2,3,4, v=1,2,…,Z), then the standardized objective function values ​​of all solutions constitute a Z-row, 4-column evaluation matrix A=[a vp ] Z×4 .

[0242] Data normalization: Since the objective function value has been standardized to the [0,1] interval and uniformly formed as a minimization form through step S500, it can be directly used for entropy calculation;

[0243] Calculate entropy and weights: Calculate the entropy e of each optimization objective.p =-(1 / lnZ)×∑ v=1 Z (P vp lnP vp ), where P vp =a vp / ∑ v=1 Z (a vp ); Calculate the objective weight w of each objective based on the entropy value. obj p =(1-e p ) / ∑ p=1 4 (1-e p The smaller the entropy value (i.e. the greater the data difference), the higher the objective weight is, reflecting its greater contribution in distinguishing the merits of different schemes.

[0244] (3) Combination weight fusion

[0245] Subjective and objective weights are combined according to a preset ratio to obtain the combined weight vector w for each optimization objective. com =[w1 com w2 com w3 com w4 com ]:

[0246] w p com =α×w sub p +(1-α)×w obj p .

[0247] Where α∈[0,1] is the subjective preference coefficient, used to balance the relative importance of expert experience and data objectivity in decision-making. α can be determined in several ways:

[0248] Based on decision-making preferences: if port planning places more emphasis on policy guidance and expert experience, α can be set to >0.5; if it places more emphasis on data-driven approaches and objective differences, α can be set to <0.5.

[0249] Determining based on data confidence: By analyzing the completeness and reliability of historical data, appropriately increase the objective weighting of targets with high data confidence.

[0250] Sensitivity analysis determined that: within the range of α values, grid search was performed to analyze the impact of different fusion ratios on the final ranking results, and an α value that made the ranking results relatively stable was selected.

[0251] (4) Calculation of comprehensive evaluation value and selection of schemes

[0252] Based on the combined weight vector, calculate the solution x for each configuration in the Pareto optimal solution set. v The overall evaluation value V v =∑ p=1 4 (w p com ×f p norm (x) v )).

[0253] Since all objective functions have been standardized to a minimization form (smaller values ​​are better), a smaller overall evaluation value indicates a better solution. The configuration solution with the smallest overall evaluation value is selected as the final configuration strategy.

[0254] The final configuration strategy is the recommended scheme for shore power berth construction that takes into account economic efficiency, environmental benefits, operational efficiency, and policy guidance. If multiple configuration solutions have similar comprehensive evaluation values ​​(with differences less than a preset threshold), a secondary screening can be conducted by combining engineering experience or additional constraints, or multiple alternative schemes can be output for decision-making reference.

[0255] Furthermore, after S600 selects the configuration solution with the highest comprehensive evaluation value, the following steps are also included:

[0256] S650 verifies the spatial balance of the configuration solution with the highest comprehensive evaluation value selected by S600 under the benchmark parameters.

[0257] Under the condition that the core uncertain parameters are taken as historical statistical values ​​or planning benchmark values, the spatial equilibrium of the configuration solution among various port areas is further examined to ensure that the final solution is not only optimal in terms of overall objectives, but also meets the preset equilibrium requirements in the spatial dimension. The specific verification process is as follows:

[0258] (1) Preset equilibrium requirements

[0259] Based on port planning and management requirements, allowable deviation ranges for three equilibrium indicators are set for each area:

[0260] Resource allocation consumption deviation: The deviation between the actual resource allocation consumption and the expected consumption of each area shall not exceed the first preset deviation value (e.g., 10%).

[0261] Shore power utilization deviation: The deviation between the actual shore power utilization rate of each area and the average utilization rate of the whole port shall not exceed the second preset deviation value (e.g., 5 percentage points).

[0262] Dedicated berth coverage deviation: The deviation between the actual dedicated berth coverage of each area and the average coverage of the entire port shall not exceed the third preset deviation value (e.g., 5 percentage points).

[0263] The aforementioned deviation thresholds can be predetermined based on engineering experience, policy requirements, or historical data statistical distribution.

[0264] (2) Balance test

[0265] Calculate the three equilibrium indices of the current optimal configuration solution in each region, and verify whether each falls within the preset allowable deviation range:

[0266] If all regions meet the requirements, the solution is determined to have passed the spatial balance verification and proceeds to step S900 for disturbance rejection capability verification.

[0267] If any indicator exceeds the allowable range, the solution is deemed to be unsatisfactory in terms of spatial balance, and parameter adjustment and re-optimization are required.

[0268] (3) Parameter tuning and re-optimization

[0269] For configuration solutions that do not meet the equilibrium requirements, adjust the optimization model parameters as follows, and re-execute steps S500 to S600:

[0270] Adjust the equilibrium penalty coefficient: Appropriately increase the equilibrium penalty coefficient of the corresponding target (e.g., increase it by a fixed step size or multiply it by a factor greater than 1) to strengthen the constraint on the equilibrium of the region during the optimization process.

[0271] Adjusting the regional balance coefficients: If the original assessment of the importance of each region is considered to be biased, the analytic hierarchy process or expert scoring method can be used again to correct the balance coefficients of each region, so that the optimization model can more accurately reflect the regional balance requirements.

[0272] (4) Iteration termination condition

[0273] Repeat the "optimize-check-adjust" process until a configuration solution that simultaneously satisfies the optimal comprehensive evaluation value (i.e., the highest comprehensive evaluation value in the current iteration) and the preset equilibrium requirement is obtained, or the preset maximum number of iterations is reached. If no solution that satisfies the equilibrium requirement is found after reaching the maximum number of iterations, the feasible solution with the best equilibrium is output and a prompt is given.

[0274] This verification step complements the "regional equilibrium stability verification" in step S700: S650 ensures that the configuration solution has spatial equilibrium under the baseline parameters, and S700 further verifies whether this equilibrium can be maintained under parameter fluctuation conditions. Through two-level equilibrium verification, the final recommended scheme combines overall optimality, spatial equilibrium, and disturbance resistance.

[0275] S700: Perform core parameter sensitivity testing and disturbance rejection capability verification on the final configuration strategy, evaluate whether the final configuration strategy still meets the constraints under parameter fluctuation conditions, and output a recommended configuration strategy that has been verified for feasibility.

[0276] Due to the uncertainties in the actual operating environment of ports, fluctuations in key input parameters may cause planning schemes to deviate from expectations or even fail. Therefore, this invention verifies the resilience of the final configuration strategy selected in step S600 to ensure its feasibility under parameter fluctuation conditions. The verification process includes two levels: basic verification and multi-condition joint verification.

[0277] (I) Identification of core uncertain parameters

[0278] Based on historical port operation data and engineering experience, key parameters that significantly impact the feasibility of the planning scheme and are subject to uncertainty were identified, including:

[0279] Resource allocation and consumption: The amount of resources required for the construction of shore power facilities at berths fluctuates due to factors such as market prices and construction conditions;

[0280] Number of berthings of vessels equipped with power receiving facilities: The actual berthing frequency may differ from historical statistics due to factors such as vessel technology updates and route adjustments.

[0281] Load factor: The actual load rate of auxiliary machinery during berthing, which is affected by factors such as the ship's operating status and season.

[0282] (II) Verification of basic disturbance immunity

[0283] Based on the identification of key uncertain parameters, the basic disturbance resistance capability of the final configuration strategy is verified:

[0284] Define the fluctuation range of parameters: Based on historical data statistics or engineering experience, set the fluctuation range of each core parameter. For example, the fluctuation range of resource allocation consumption is ±10%, the fluctuation range of the number of times ships with power receiving facilities berth is ±15%, and the fluctuation range of load factor is 0.4~0.6.

[0285] Construct the worst-case scenario: In the worst-case scenario, all parameters are simultaneously taken to the values ​​most unfavorable to the configuration strategy—even the parameter combination with the highest resource consumption, lowest emission reduction benefits, and lowest utilization rate, to test whether the configuration strategy still meets the constraints.

[0286] Verify constraint satisfaction: In the worst-case scenario, recalculate the upper limit constraint of total resources, berth type constraint, and spatial distribution constraint. If the final configuration strategy still satisfies all constraints, the basic disturbance resistance capability is verified; otherwise, the configuration strategy needs to be adjusted or re-optimized.

[0287] (III) Multi-condition joint verification

[0288] To more comprehensively evaluate the feasibility of the configuration strategy under real-world uncertainty, this invention further introduces multi-condition joint verification based on scenario clustering:

[0289] Generating parameter fluctuation scenarios: Based on historical data, a joint distribution model of core uncertain parameters is constructed. A large number (e.g., 10,000) of parameter fluctuation scenarios are generated using the Monte Carlo sampling method. Each parameter fluctuation scenario corresponds to a set of core parameter value combinations. The historical data comes from the multi-source fusion database constructed by S100, particularly the ship-berth berthing information table, planning guidance technical parameter table, and sensitivity analysis parameter table. By extracting the accumulated operational records and statistical information from these data tables over the years, a sample set of core uncertain parameters is formed, providing a data foundation for the construction of the joint distribution model.

[0290] Scene clustering reduction: Clustering algorithms (such as K-means or DBSCAN) are used to cluster the generated parameter fluctuation scenarios into several typical working condition scenarios. Each typical working condition scenario includes:

[0291] Scene parameter values: The combination of parameters corresponding to the cluster centers;

[0292] Scenario occurrence probability: The proportion of the number of scenarios included in this cluster out of the total number of scenarios.

[0293] Multi-condition constraint satisfaction check: Under each typical operating condition, verify whether the final configuration strategy satisfies all constraints, and calculate the probability of constraint satisfaction, Psatisfy: Psatisfy = ∑ g=1 Q (p) g ×I g Where Q represents the number of typical working scenarios, and p g Let I be the probability of the g-th typical working condition scenario occurring. g This is an indicator function (1 for all constraints in this typical scenario, 0 for others).

[0294] Damage resistance assessment: A preset damage resistance threshold Pth (e.g., 95%) is used. If Psatisfy ≥ Pth, it indicates that the final configuration strategy remains feasible under most possible real-world operating conditions, and the final configuration strategy is deemed to have damage resistance. Otherwise, the configuration strategy needs to be adjusted or re-optimized until the threshold requirement is met. For typical scenarios that do not meet the constraints (I... g If the value is 0, its parameter characteristics can be further analyzed to identify the key parameter combinations that cause the solution to fail, providing a basis for subsequent solution adjustments.

[0295] (iv) Verification of regional equilibrium stability

[0296] Based on multi-condition joint verification, this invention further verifies the regional equilibrium stability of the final configuration strategy under parameter fluctuations, ensuring that it can maintain balanced development among regions even in uncertain environments:

[0297] Define the regional equilibrium stability index: For each typical operating scenario, calculate the three equilibrium indices for each region according to the same equilibrium index definition as in S850:

[0298] Resource allocation consumption deviation: The absolute value of the difference between the sum of the actual resource allocation consumption of all berths equipped with shore power in this area and the expected resource allocation consumption of this area in this scenario;

[0299] Shore power utilization deviation: The absolute value of the difference between the actual shore power utilization rate of this area and the average shore power utilization rate of all areas in Hong Kong under this scenario;

[0300] Dedicated berth coverage deviation: The absolute value of the difference between the actual dedicated berth coverage of this area and the average dedicated berth coverage of all areas in Hong Kong under this scenario.

[0301] The criteria for determining balance and stability are set as follows: the allowable deviation range preset in S650 is used as the criterion, namely, the deviation of resource allocation consumption does not exceed the first preset deviation value, the deviation of shore power utilization rate does not exceed the second preset deviation value, and the deviation of dedicated berth coverage rate does not exceed the third preset deviation value. If the deviations of all three items in a certain scenario do not exceed the corresponding thresholds, then the scenario is considered to meet the area balance and stability requirements.

[0302] Calculate the equilibrium stability probability: Sum the probabilities of occurrence of all typical operating conditions that meet the requirements to obtain the equilibrium stability probability Pbalance.

[0303] Balance stability determination: A preset balance stability probability threshold Pbalance_th (e.g., 90%) is set. If Pbalance ≥ Pbalance_th, it is determined that the final configuration strategy can still maintain the balance stability of the region under parameter fluctuations; otherwise, the balance penalty coefficient or the region balance coefficient needs to be adjusted and the optimization process needs to be re-executed.

[0304] (v) Recommended configuration strategy output

[0305] After verifying the above-mentioned resilience to disturbances, the final configuration strategy and its resilience assessment results (including constraint satisfaction under worst-case scenarios, constraint satisfaction probabilities under multiple operating conditions, sensitive parameters and their impact) will be output together as a recommended configuration strategy with verified feasibility for port planning decisions. This output ensures that the recommended shore power berth construction scheme is not only optimal under ideal conditions, but also robustly feasible under actual fluctuating operating conditions.

[0306] Furthermore, after selecting the configuration solution with the highest comprehensive evaluation value in step S600, the spatial balance of the solution needs to be further verified to ensure that the final solution is not only optimal in terms of the overall objective, but also meets the preset balance requirements among the various port areas. Specifically, this includes:

[0307] Preset equilibrium requirements: Based on port planning and management needs, set allowable deviation ranges for resource allocation consumption, shore power utilization, and dedicated berth coverage in each area. For example, the deviation between actual resource allocation consumption and expected consumption in each area should not exceed a first preset deviation value (e.g., 10%), the deviation between shore power utilization in each area and the port-wide average utilization rate should not exceed a second preset deviation value (e.g., 5 percentage points), and the deviation between dedicated berth coverage in each area and the port-wide average coverage rate should not exceed a third preset deviation value (e.g., 5 percentage points). These thresholds can be predetermined based on engineering experience or policy requirements.

[0308] Spatial balance test: Calculate the three indicators mentioned above for the current optimal configuration solution in each region, and check whether each indicator falls within the preset allowable deviation range. If all regions meet the requirements, the solution is confirmed as the final recommended configuration strategy; if any indicator exceeds the allowable range, the solution is determined to fail in terms of spatial balance.

[0309] Parameter tuning and re-optimization: For solutions that do not meet the equilibrium requirements, adjust the optimization model parameters in the following ways and re-execute steps S500 to S600:

[0310] Adjust the equilibrium penalty coefficient: Appropriately increase the equilibrium penalty coefficient of the corresponding target (e.g., increase it by a fixed step size or multiply it by a factor greater than 1) to strengthen the constraint on the equilibrium of the region during the optimization process.

[0311] Adjusting the regional balance coefficients: If the original assessment of the importance of each region is considered to be biased, the analytic hierarchy process or expert scoring method can be used again to correct the balance coefficients of each region, so that the optimization model can more accurately reflect the regional balance requirements.

[0312] Iteration termination condition: Repeat the above "optimization-verification-adjustment" process until a configuration solution that simultaneously satisfies the optimal comprehensive evaluation value (i.e., the highest comprehensive evaluation value in the current iteration) and the preset equilibrium requirement is obtained, or the preset maximum number of iterations is reached. If no solution that satisfies the equilibrium requirement is found after reaching the maximum number of iterations, the feasible solution with the best equilibrium is output and a prompt is given.

[0313] This additional verification step ensures that the final recommended scheme not only performs well in the four optimization objectives, but also achieves a balanced allocation of resources in the spatial dimension, avoiding the problems of excessive concentration of resources or unbalanced development in the area, and improving the practical acceptability of the planning scheme.

[0314] This invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to perform the method described in this invention.

[0315] This invention also provides a computer-readable storage medium storing computer-executable instructions for performing the methods described in this invention.

[0316] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0317] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A multi-objective collaborative optimization planning method for the construction of shore power berths in ports, characterized in that, The method includes the following steps: S100, acquire multi-source fusion data, which includes berth basic information, ship-berth berthing correlation data, planning guidance technical parameters, and sensitivity analysis parameters; S200, based on the multi-source fusion data, binary decision variables are used to characterize whether each berth is equipped with shore power facilities; S300 constructs at least several optimization objective functions, including minimizing resource consumption, maximizing carbon emission reduction benefits, maximizing shore power utilization, and maximizing planning guidance adaptability. The construction logic of the multiple optimization objective functions is as follows: The objective function for minimizing resource consumption is: ; where x j Let C be the binary decision variable for the j-th berth. j The resource allocation cost for shore power facilities at the j-th berth, where n is the total number of port berths; j ranges from 1 to n. The objective function for maximizing carbon emission reduction benefits is: ; Where m is the number of ship types, SP ij T represents the number of times a vessel of class i equipped with power receiving facilities berths at berth j. ij Let P be the average berthing time per trip for vessel of class i. i For the rated power of the auxiliary engine of the i-th type of ship, ρ i For load factor, For fuel consumption rate, EF i CO2 emission factor for fuel used in auxiliary equipment; i ranges from 1 to m; The objective function for maximizing shore power utilization is: ; Among them, Q ij The total number of times a vessel of type i berths at the j-th berth; The objective function for maximizing planning-oriented adaptability is: Where S is the set of dedicated berths, and |S| is the total number of dedicated berths; S400, based on the multi-source fusion data, the pre-set constraints of the multiple optimization objective functions, and the standardized objective function values, a multi-objective optimization model is constructed; wherein, the constraints are used to limit the feasible solution space for shore power facility configuration, and the standardization process is used to map objective values ​​of different dimensions to a unified scale; the pre-set constraints specifically include: Total resource limit constraint: Among them, B max This represents the upper limit of the total available resources for shore power construction; Berth type constraints: According to the requirements for shore power promotion, the construction of shore power must be provided for dedicated berths or general-purpose berths, or shore power construction is prohibited. Spatial distribution constraints: At least one shore power berth must be constructed in each port area; Decision variable constraints: x j ∈{0,1}; S500 employs an improved NSGA-Ⅲ algorithm to solve the multi-objective optimization model within the feasible solution space defined by the constraints, thereby obtaining the Pareto optimal configuration solution set. The improved NSGA-III algorithm introduces an adaptive mutation mechanism and a dynamic reference point adaptive adjustment mechanism. The adaptive mutation mechanism dynamically adjusts the mutation probability according to the iteration process, so that the mutation probability decreases as the number of iterations increases, and the crowding distance method is used to maintain the uniformity of the Pareto solution set distribution. Non-dominated solutions are stored in an external archive and periodically pruned. The dynamic reference point adaptive adjustment mechanism adjusts the position and density of reference points in real time according to the distribution status of each generation of the population. Specifically, it includes: calculating the distribution boundary and aggregation degree of the current population in the target space; dynamically shrinking or expanding the coverage of reference points based on the distribution boundary so that the reference points are always distributed in the effective area of ​​the current Pareto front; automatically interpolating to generate new reference points in sparse regions of the solution set based on the aggregation degree, and merging neighboring reference points in dense regions to achieve adaptive adjustment of the reference point density. S600, calculate the comprehensive evaluation value of each configuration solution in the Pareto optimal configuration solution set based on the combined weight method, and select the configuration solution with the highest comprehensive evaluation value as the final configuration strategy for the construction of shore power berths; S700: Perform core parameter sensitivity testing and disturbance rejection capability verification on the final configuration strategy, evaluate whether the final configuration strategy still meets the constraints under parameter fluctuation conditions, and output a recommended configuration strategy that has been verified for feasibility.

2. The method according to claim 1, characterized in that, The multi-source fusion data is organized and managed in the form of database tables, specifically including: The berth basic information table includes fields such as berth identifier, berth type, berth tonnage, specialized berth identifier, shore power facility resource allocation and consumption, total number of berths per year, and port area to which it belongs. The vessel-berth berthing information table includes the berth identifier, vessel type, number of berthings per year for vessels with power receiving facilities, average berthing time per berthing, and auxiliary engine power of the berthed vessel. The planning and technical parameters table includes the upper limit of the total available resources for shore power construction, the planning target for shore power coverage of specialized berths, the load factor, and the CO2 emission factor fields. The sensitivity analysis parameter table includes fields for parameter type, parameter name, baseline value, range of change, and value after change.

3. The method according to claim 1, characterized in that, The combined weighting method is an entropy weight-fuzzy hierarchical analysis method, specifically including: The subjective weights of each optimization objective function are calculated using the fuzzy hierarchical analysis method. The expert fuzzy judgment is processed by triangular fuzzy numbers, and the subjective weight vector is obtained after defuzzification and normalization. The objective weights of each optimization objective function are calculated using the entropy weight method, and the objective importance of each optimization objective function is determined based on the information entropy of historical data. Subjective weights and objective weights are combined according to a preset ratio to obtain the combined weights of each optimization objective function.

4. The method according to claim 1, characterized in that, The disturbance rejection capability verification specifically includes: Identify the core uncertain parameters, which include resource allocation consumption, the number of times ships equipped with power receiving facilities berth, and the load factor; Define the fluctuation range or uncertainty set of each core uncertainty parameter; The deterministic constraints are transformed into disturbance resistance constraints, and the final configuration strategy is tested to see if it still satisfies the upper limit of total resources, berth type constraints, and decision variable constraints under the worst-case scenario of parameter fluctuations.

5. An electronic device, characterized in that, Including processor and memory; The processor executes the steps of the method as described in any one of claims 1 to 4 by invoking programs or instructions stored in the memory.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a program or instructions that cause a computer to perform the steps of the method as described in any one of claims 1 to 4.

Images