Hierarchical Optimization Method Research Articles

Hierarchical optimization of reservoir development strategy based on reinforcement learning

Production optimization is to increase the profitability of reservoir development by adjusting the type of wells and the rate of injection and production. Previous methods put both variables at the same level when optimizing and thus they change at the same frequency. However, these two variables belong to different scales, because well types require time-consuming and labor consuming operations such as borehole operations, while rate adjustment don't. Therefore, in this paper, a hierarchical optimization method is proposed to optimally control well types and corresponding rate in production sequence. Specifically, a hierarchical framework is first employed based on reinforcement learning method, with well types as high-level variables and quantities as low-level variables. Then, two hindsight mechanisms, named action hindsight and goal hindsight, are used to achieve stably joint optimization of two-level variables at different frequencies. Our findings from the case study reveal that our approach is highly practical, as it has successfully reduced operational costs by 90% and increased cumulative oil production by 18% within three years when compared to the single-level reinforcement method.

Hierarchical optimization method of space patrol task assignment based on deep neural network and consensus-based bundle algorithm

This research looks towards improving the accuracy and efficiency of orbit transfer speed increment estimation and task assignment for space patrol missions involving multiple platforms. First, we set up a quick prediction model for the increment in orbit transfer velocity. Then, a two-layered hybrid optimization framework is developed, with a deep neural network sub-model as the inner layer to find the optimal speed increase and the consensus-based bundle algorithm as the outer layer to solve the multi-task patrol sequence problem. Finally, the simulation and analysis of the patrol task assignment problem are conducted under two types of scenarios. The results show that the hierarchical optimization method proposed in this paper is three orders of magnitude faster than the traditional optimization method. In addition, the suggested hierarchical optimization method has significant promise in practical applications, where it can produce solutions more quickly to adapt to more complex patrol scenarios.

Quasi-perfect and wideband absorption with gradient non-uniform micro-slit array absorber via a hierarchical optimization method

Gradient micro-slit array absorber, a kind of acoustic metamaterial, made of multiple quarter-wavelength resonators, is holding a promising future in industrial utilization due to its flexible adaptation and simple structure. However, designing a perfect and broadband absorber remains a huge challenge because of the larger number of the design parameters and complicated coupling effect between different resonators. In this study, a hierarchical optimization method is developed for gradient micro-slit array absorber. First, a prior test based on two-slit absorber is conducted using transfer matrix method (TMM) to give directions on designing a multi-slit absorber. And then an original absorber model is constructed in the range of 1000–1800 Hz. For the first step, the optimal width of the slit is obtained through the analysis of the quality factor. For the second step, the spacings between slits is optimized by maximizing absorption and minimizing the total length of the waveguide. After optimization, the absorption performance of the optimized model increased by 15.6% and the length decreased by 46% compared with the original model. Finally, the gradient micro-slit array absorber is experimentally investigated and achieves a quasi-perfect and flat absorption ranging from 1000 Hz to 1800 Hz.

Automated and Clinical-Criteria-Driven Planning for Lung Cancer Using the Expedited Constrained Hierarchical Optimization (ECHO) System

<h3>Purpose/Objective(s)</h3> Treatment planning for locally advanced non-small cell lung cancer (NSCLC) has always been challenging due to the variabilities of the planning target volume (PTV) and competing constraints of organs at risk such as lung, cord, esophagus, and heart. We share our experience in developing a fully automatic treatment planning solution for NSCLC. <h3>Materials/Methods</h3> Automated plans for NSCLC radiotherapy patients with static IMRT fields were designed using an in-house-developed automated planning system named Expedited Constrained Hierarchical Optimization (ECHO) which is integrated into the treatment planning system (TPS) using API scripting. ECHO does not use TPS optimization algorithms, but utilizes TPS leaf sequencing and final dose calculation. ECHO applies advanced optimization tools (such as hierarchical constrained optimization, convex approximations and Lagrangian methods) to produce Pareto optimal plans that meet clinical constraints and improve desired clinical objectives as much as possible. ECHO optimization parameters were tuned with 29 NSCLC patients and validated retrospectively using 113 NSCLC patients treated with prescription doses including conventional (1.8Gy-2Gy x 25-35 fractions) and hypofractionation (2.75Gy-4Gy x 10-20 fractions). Clinically treated plans were created manually using either static IMRT or VMAT. Automated plans were created using static IMRT fields with beam templates based on the location of PTV within the lung. Dosimetric metrics were evaluated with respect to tumor coverage and normal tissue sparing. <h3>Results</h3> PTV sizes ranged from 64cc to 2297cc (median 292cc). Median planning time for ECHO was 23 minutes (range: 8min-78min). Both manual and ECHO plans met all established clinical criteria. Compared with manual plans, ECHO plans produced significantly improved and more consistent PTV coverage and better sparing of esophagus, heart and cord while maintaining similar lung doses (Table 1). <h3>Conclusion</h3> An automated and clinical-criteria-driven treatment planning system has been validated and clinically implemented in our institution for NSCLC radiotherapy. Use of automation improves efficiency, plan consistency and quality for NSCLC treatment planning.

A multi‐priority hierarchical optimization method for double‐layer model predictive control

AbstractConsidering the demand for the sequential regulation of manipulated variables in actual industrial process control, the conventional solution of double‐layer model predictive control faces the problem that the weight coefficients are difficult to tune. This paper proposes an improved hierarchical optimization method for manipulated variables in the steady‐state optimization layer of double‐layer model predictive control. The proposed method can adjust the manipulated variables sequentially without an accurate weight coefficient to avoid difficulty in tuning the weight coefficients. The relation between the optimal solution and the feasible region of the steady‐state optimization layer is analysed to describe the reoptimization of the key manipulated variables. The impact of the economic cost coefficient on the optimal solution with the sensitivity analysis method is studied, and the complexity of using the weight coefficient to solve the priority optimization problem of the manipulated variables is assessed. The steady‐state optimization solution procedure is improved based on the theory of the multiobjective complete hierarchical method. The hierarchical and sequential optimization of the manipulated variables results in expanding the space and freedom of the key manipulated variables, increasing efficiency, reducing consumption, and improving economic performance. The improved hierarchical optimization method is direct and simple in achieving optimization sequentially and satisfies the need for adjusting the manipulated variables according to human intentions.

Research on the utilization of fuel cold energy for LNG powered container ships

Aiming at the problems that the fuel releases a lot of cold energy and the refrigerated containers consume a lot of electricity on large LNG powered container ships, a set of cold energy cascade utilization scheme that mainly uses LNG cold energy for refrigerated containers is designed. The simulation software Aspen HYSYS is used to establish and simulate the process of the ship’s cold energy cascade utilization system under five different working conditions, and the main parameters of the key nodes are obtained, according to the established exergy efficiency model, the exergy efficiency of the main equipment and the whole system is solved. Under the five working conditions, the maximum exergy efficiency of the refrigerated container system before optimization is 24.54%, at this time, the exergy efficiency of the entire cold energy utilization system is 24.86%. With the goal of improving the exergy efficiency of the entire system, using a hierarchical optimization method, the key parameters affecting the exergy efficiency of each cold energy utilization module are analyzed and optimized respectively, and the optimal operating parameters of different cold energy utilization module were determined. So as to realize the optimization of the whole cold energy cascade utilization system. The results show that the exergy efficiency of the optimized LNG cold energy cascade utilization system for ships under five working conditions is improved, under the 65% working condition, the exergy efficiency of the optimized refrigerated container system is 27.69%, and the exergy efficiency of the whole cold energy utilization system is 28.04%, which are increased by 3.15% and 3.18% respectively. Which proves that the LNG cold energy cascade utilization system can realize the effective utilization of LNG cold energy.

A hierarchical approach for building distributed quantum systems

In this paper, a multi-layer hierarchical architecture is proposed for distributing quantum computation. In a distributed quantum computing (DQC), different units or subsystems communicate by teleportation in order to transfer quantum information. Quantum teleportation requires classical and quantum resources and hence, it is essential to minimize the number of communications among these subsystems. To this end, a two-level hierarchical optimization method is proposed to distribute the qubits among different parts. In Level I, an integer linear programming model is presented to distribute a monolithic quantum system into K balanced partitions which results in the minimum number of non-local gates. When a qubit is teleported to a destination part, it can be used optimally by other gates without being teleported back to the destination part. In Level II, a data structure is proposed for quantum circuit and a recursive function is applied to minimize the number of teleportations. Experimental results show that the proposed approach outperforms the previous ones.

Position‐track‐time three‐dimensional network model‐based improvement for high‐speed rail scheduling and operation

AbstractThe train timetable and station operation plan play a critical role in the high‐speed railway (HSR) planning and management. The existing hierarchical optimization methods for the planning process of the HSR would affect the efficiency of train schedules and are often difficult to achieve an optimized scheme. This paper proposes a position‐track‐time three‐dimensional network, which describes the process of train operations in sections and stations at a macroscopic scale, while the track infrastructure including the position of insulation joints in stations are modeled microscopically. The modeled train running times and dwell times are based on standard timetable design values given in full minutes by the China Railway Corporation, while the interlocking times and minimum headway times are not specified explicitly. The problem is formulated as a large‐scale 0–1 linear integer programming model, which is solved using an extended branch‐and‐price algorithm. The effectiveness and precision of the model are verified through a real‐world case study on the Beijing–Shanghai HSR line. The results indicate that the proposed model can effectively improve the line capacity by 17.2% while ensuring that there is no conflict between train operations in sections and stations.

Hierarchical Scheduling Control Method for Cascade Hydro-PV-Pumped Storage Generation System

High photovoltaic penetration in a power system has significantly challenged its safety and economic operation. To use the complementary characteristics of various renewable energy sources (RESs) fully, a novel hierarchical scheduling control (HSC) method is presented to accommodate the variability and uncertainty of a cascade hydro-PV-pumped storage (CH-PV-PS) generation system. Considering the optimization functions and execution requirements of the CH-PV-PS system, the HSC method is divided into two layers: the dynamic optimization layer and the static optimization layer. The static optimization layer focuses on the economy of the CH-PV-PS system, and the dynamic optimization layer focuses on the safety of the CH-PV-PS system. In the first layer, that is, the static optimization layer, the objectives of the day-ahead and hour-ahead schedules are established, and a heuristic algorithm is combined with a linear programming algorithm to optimize the energy allocation. Considering the uncertainty of the PV power output and hour-ahead load, a real-time schedule is established in the second layer; that is, in the second layer, the dynamic optimization layer, real-time scheduling and prediction of active output are established. Model predictive control methods are introduced to correct for prediction bias at different time scales in order to fully utilize the control capability of hydropower generation. A CH-PV-PS real-world system in Southwest China is chosen as a case study. In the three scenarios, where only PV fluctuations are considered, the simulation results reveal that, compared with the traditional open-loop optimized and hierarchical open-loop optimization methods, the HSC method reduces the average relative deviation of PV and increases the system economics. After a large amount of RESs are connected to the power grid, the HSC method provides a solution for improving the consumption of RESs.

A Hierarchical Optimization Method for the Design of Active Hybrid Structures

In structural engineering, active structures that combine the principles of lightweight construction with bending elastic component behavior are increasingly being investigated. For the realization of a prototype of an active hybrid roof structure at the laboratory of Hybrid Structures at BTU Cottbus-Senftenberg, preliminary investigations on a case study are conducted in the framework of this publication in order to improve the design process of these types of structures. These active hybrids require a higher design effort than classical structures from the field of structural engineering due to a larger number of relevant objectives. Consequently, this study devotes special attention to these essential target criteria and their mathematical formulation. Furthermore, in order to improve the efficiency of this design process, a hierarchical method is derived that is subdivided into two successive partial procedures, which contain specific heuristics that are developed. In this method, after structural optimization, an optimal actuator placement is performed. The subject of a design process involving optimal actuator placement is relatively unexplored for active structures in which components are subjected to large elastic bending deformations and is therefore the focus of this study. In order to verify the functionality of the method and the plausibility of the results of the derived partial methods, a validation of the methodology is performed. Therefore, results of analyses of an active truss structure are compared with those of an active hybrid structure, both derived using the presented method. In addition to validating results, the study intends to investigate whether the performance of an active hybrid structure generated by the proposed method is sufficiently competitive compared to a state-of-the-art active truss structure derived by the same procedure.

Hypergraph-based joint optimization of spare part provision and maintenance scheduling for serial-parallel multi-station manufacturing systems

This paper studies the joint modeling of spare part provision and maintenance scheduling for Serial-Parallel Multi-station Manufacturing Systems (SPMMS), where the spare part connections are especially investigated. In the SPMMS, multiple types of spare parts are retained for uncertain failures of the stations, and multiple stations may share the spare parts in inventory. To describe the spare part connections among the stations, a hypergraph-based model is introduced. In the system-level maintenance, the SPMMS is formulated to be a multi-connection hypergraph, with the Spare Part Hyperedges (SPH), the cost-saving hyperedges and the capacity-loss hyperedge involved. Based on the multi-connection hypergraph, a Combination-Spread Opportunistic Maintenance (CSOM) policy is proposed. To obtain the optimal alternatives of the maximum inventory levels, the length of the combination window and the length of the spread window, a hierarchical optimization method is developed. The case study and the comparisons show that the proposed hypergraph-based CSOM policy is more cost-effective than the traditional maintenance policy, the station-level maintenance policy and the fixed-group maintenance policy. Finally, the discussion shows that the traditional maintenance model is a special scenario of the hypergraph-based model.

Hydrodynamic interactions among wave energy converter array and a hierarchical genetic algorithm for layout optimization

This paper focuses on the layout optimization and hydrodynamic investigation of wave energy converter (WEC) array. For hydrodynamic analysis, each buoy is modeled as a truncated floating cylinder that can move with five degrees of freedom. The hydrodynamic response of such an array is analyzed by a semi-analytical method. Then, the power extraction by the array is used as the objective function for layout optimization. To improve efficiency, especially for an array with a large number of WECs, a novel hierarchical optimization method based on matrix coded genetic algorithm is proposed. Within this method, a large-scale array is firstly divided into several levels of small-scale arrays which are easy to handle, then, the genetic algorithm developed in this paper is used to optimize the small array of each level. An example of two-level optimization is given in this paper and compared with the result by conventional optimization routine. Finally, we investigate the influence of introducing other degrees of freedom, besides the heave mode, on the performance of arrays. Results show that as the constraints get weaker in non-heave directions, the effect of other degrees of freedom on the WEC array interaction factor becomes more and more obvious.

Genetic variability in Capsicum spp. accessions through multicategorical traits

Capsicum peppers originate from the American continent. Their culinary and medicinal potential, first exploited by indigenous peoples over 7000 years ago is vast, however more studies are needed. Despite the existence of germplasm banks that house hundreds of accessions of the genus, only a little of this diversity is seen in large markets. In this respect, the morpho-agronomic characterization of the accessions allows not only the identification of the existing variability, but also the agronomic and commercial potential of the genotypes. Therefore, this study proposes to characterize and estimate genetic divergence between accessions of the Capsicum germplasm collection at the Universidade Federal Rural do Rio de Janeiro (UFRRJ) based on multicategorical traits. The species were identified and 29 accessions were characterized in a greenhouse, based on 31 multicategorical descriptors of the crop. The collected data were subjected to the Cole-Rodgers dissimilarity index and to hierarchical clustering (Nearest Neighbor) and optimization (Tocher’s) methods. Tocher's method was considered the most suitable to group the accessions of the germplasm bank. Results indicate that the multicategorical descriptors properly assess the diversity in the chosen method and display the existence of high variability in the collection, which can be used in Capsicum spp. breeding programs.

Accelerated hierarchical optimization method for emergency energy management of microgrids with energy storage systems

AbstractWhen failures occur in microgrids (MGs), the energy management for emergencies is required. To respond to emergencies in MGs rapidly, an accelerated hierarchical optimization method has been proposed, where the outputs of energy storage systems (ESSs) are controlled to provide urgent supports, before the MG reconfiguration starts. However, it is time‐consuming to find the optimal schemes for MG reconfiguration. To reduce the time of reconfiguration, an optimization method based on deep neural networks (DNNs) for MG reconfiguration is presented, which consists of three levels. First, the combinational optimization for load shedding amounts is solved to determine the loads that have to be cut off. Second, a DNN is used to find the parameters of reconfiguration instead of the time‐consuming calculation of power flow, which is a good way to reduce time of reconfiguration. Third, according to these parameters, the optimal scheme for reconfiguration is selected by a comprehensive evaluation method, where the Delphi method (DM) is employed to adjust the weights of preferences in a comprehensive evaluation function, so it offers the diversity of decisions for the MG reconfiguration. Finally, to test our method, a modified IEEE 33‐bus system is built in MATLAB for simulations. Compared to traditional methods, our method can obtain the same reconfiguration scheme under different on/off states of load switches, but the time of reconfiguration is only one‐sixty‐seventh of that of other methods. Furthermore, in terms of our comprehensive evaluation method, reconfiguration schemes can be selected under different preferences.

Integrated optimization design of electric power steering and suspension systems based on hierarchical coordination optimization method

Integrated optimization design of electric power steering and suspension systems based on hierarchical coordination optimization method

Fundamental Study of Hierarchical Optimization Methods for Product Systems Using Reinforcement Learning

As product systems become larger and more complex, their design space is increasing, and it is becoming more difficult to optimize them. In the past, hierarchical optimization methods have been proposed to solve this problem by decomposing the entire system into several subsystems and dividing and reducing the design space. However, they are ineffective for problems with strong dependencies among many subsystems. On the other hand, it has become clear that reinforcement learning can be used to efficiently search for optimal solutions to optimization problems such as solving small-scale games. If we try to apply reinforcement learning to large-scale optimization problems, the learning space of reinforcement learning increase, and optimization become difficult. Therefore, in this study, we used multi-agent reinforcement learning to reduce the learning space handled by reinforcement learning at a time. In this method, we propose an algorithm that reduces the learning space by allowing each agent to learn only the priority of a subsystem based on the relationship between the design variables and the evaluation values of each subsystem. The proposed method reduced the number of evaluations to less than 10% of the conventional number while maintaining the solution quality of the optimization.

Automated and Clinically Optimal Treatment Planning for Cancer Radiotherapy.

Each year, approximately 18 million new cancer cases are diagnosed worldwide, and about half must be treated with radiotherapy. A successful treatment requires treatment planning with the customization of penetrating radiation beams to sterilize cancerous cells without harming nearby normal organs and tissues. This process currently involves extensive manual tuning of parameters by an expert planner, making it a time-consuming and labor-intensive process, with quality and immediacy of critical care dependent on the planner's expertise. To improve the speed, quality, and availability of this highly specialized care, Memorial Sloan Kettering Cancer Center developed and applied advanced optimization tools to this problem (e.g., using hierarchical constrained optimization, convex approximations, and Lagrangian methods). This resulted in both a greatly improved radiotherapy treatment planning process and the generation of reliable and consistent high-quality plans that reflect clinical priorities. These improved techniques have been the foundation of high-quality treatments and have positively impacted over 4,000 patients to date, including numerous patients in severe pain and in urgent need of treatment who might have otherwise required longer hospital stays or undergone unnecessary surgery to control the progression of their disease. We expect that the wide distribution of the system we developed will ultimately impact patient care more broadly, including in resource-constrained countries.

Two-level optimization approach to tree-level forest planning

Laser scanning and individual-tree detection are used increasingly in forest inventories. As a consequence, methods that optimize forest management at the level of individual trees will be gradually developed and adopted. The current study proposed a hierarchical two-level optimization method for tree-level planning where the cutting years are optimized at the higher level. The lower-level optimization allocates the trees to the cutting events in an optimal way. The higher-level optimization employed differential evolution whereas the lower-level problem was solved with the simulated annealing metaheuristic. The method was demonstrated with a 30 ​m ​× ​30 ​m sample plot of planted Larix olgensis . The baseline case maximized the net present value as the only management objective. The solution suggested heavy thinning from above and a rotation length of 62 years. The baseline problem was enhanced to mixed stands where species diversity was used as another management objective. The method was also demonstrated in a problem that considered the complexity of stand structure, in addition to net present value. The objective variables that were used to measure complexity were the Shannon index (species diversity), Gini index (tree size diversity), and the index of Clark and Evans, which was used to describe the spatial distribution of trees. The article also presents a method to include natural advance regeneration in the optimization problem and optimize the parameters of simulated annealing simultaneously with the cutting years. The study showed that optimization approaches developed for forest-level planning can be adapted to problems where treatment prescriptions are required for individual trees.

Hierarchical variable fidelity evolutionary optimization methods and their applications in aerodynamic shape design

This paper proposes two hierarchical evolutionary optimization methods based on variable fidelity analysis and search space contraction for aerodynamic shape design, i.e. hierarchical evolutionary Pareto and Nash games. One of these techniques is used in the optimization process, namely the advantages of high and low fidelity flow simulation. The high-fidelity model provides solution accuracy while the low-fidelity model reduces the computational cost. Especially, the search space contraction and the population size reduction are introduced in the process of transition from the optimization on low fidelity simulation to the optimization on high fidelity simulation, so that the optimization based on high fidelity simulation can get the high-precision optimal solution with a relatively less Central Processing Unit(CPU) cost. They are applied to the single objective natural laminar wing shape design at transonic flow and the multidisciplinary shape optimization of a hypersonic air-breathing vehicle respectively. The optimization results show that regardless of a single objective or multi-objective/multidisciplinary optimization problem, the new hierarchical optimization methods proposed in this paper can improve the optimization efficiency by 5-10 times.

A Medium- and Long-Term Orderly Charging Load Planning Method for Electric Vehicles in Residential Areas

Due to the large number of electric vehicles (EVs) connected to the distribution network of residential areas (RAs), community charging has become a major constraint. The planning of the distribution network in RAs needs to consider the orderly charging load of EVs. In the current study, an orderly charging planning method for the charging posts and distribution network of RAs was proposed. First, a charging load forecasting model based on the travel characteristics, charging time, and ownership of EVs in RAs was established. Then, a hierarchical orderly charging optimization method, including a distribution network layer and EV access node layer, was devised. The upper layer optimizes the distribution network. The objective function is the minimum variance of the overall load in the RA and the constraint conditions satisfy the overall charging load demand and the capacity of the distributed network. The lower layer optimizes the EV access nodes. The objective function is the minimum variance of the node access load, and the constraint conditions are to meet the regional charging load demand and the optimal power balance demand transmitted from the upper layer to the lower layer. A nonlinear optimization algorithm is employed to solve these objective functions. An IEEE 33 node example was used to obtain the orderly charging power load curves for weekdays and weekends in RAs, and the simulation results prove the effectiveness of the proposed method.