Planning in partially observable domains is a notoriously difficult problem. However, inmany real-world scenarios, planning can be simplified by decomposing the task into a hierarchy of smaller planning problems. Seve...
详细信息
ISBN:
(纸本)9780262195683
Planning in partially observable domains is a notoriously difficult problem. However, inmany real-world scenarios, planning can be simplified by decomposing the task into a hierarchy of smaller planning problems. Several approaches have been proposed to optimize a policy that decomposes according to a hierarchy specified a priori. In this paper, we investigate the problem of automatically discovering the hierarchy. More precisely, we frame the optimization of a hierarchical policy as a non-convex optimization problem that can be solved with general non-linear solvers, a mixed-integer non-linear approximation or a form of bounded hierarchical policy iteration. By encoding the hierarchical structure as variables of the optimization problem, we can automatically discover a hierarchy. Our method is flexible enough to allow any parts of the hierarchy to be specified based on prior knowledge while letting the optimization discover the unknown parts. It can also discover hierarchical policies, including recursive policies, that are more compact (potentially infinitely fewer parameters) and often easier to understand given the decomposition induced by the hierarchy.
In this paper, the problem of finding the optimal topological configuration of a power transmission system is considered with the aim of providing system operators with a tool suited for congestion management. Network...
详细信息
In this paper, the problem of finding the optimal topological configuration of a power transmission system is considered with the aim of providing system operators with a tool suited for congestion management. Network reconfiguration looks particularly appealing since it allows transmission system operators to alleviate overloads by means of switching operations that may avoid costly generation or load curtailments. The techniques of corrective switching proposed in the 1980s are profitably employed to formulate the problem of network reconfiguration for the purpose of congestion management. The solution of the resulting large-scale mixed-integerprogramming problem is carried out both by a deterministic branch- and-bound algorithm included in the CPLEX optimization package and by a genetic algorithm. Tests were performed on a 33-bus CIGRE test system and on an actual 432-bus network of Italian origin. (c) 2005 Elsevier B.V. All rights reserved.
This paper presents an optimization based algorithm to solve the weekly scheduling problem of a large-scale hydrothermal power system, formulated as a mixed-integer linear programming model (MILP). The main drawback o...
详细信息
This paper presents an optimization based algorithm to solve the weekly scheduling problem of a large-scale hydrothermal power system, formulated as a mixed-integer linear programming model (MILP). The main drawback of the MILP approach is the high computational burden required to solve large-size problems. The proposed algorithm tackles this problem by providing an initial feasible and integer solution, which enhances the search of the Branch and Bound (B&B) over the space of feasible solutions, reducing the resolution time. A detailed representation of thermal, pumped storage, and hydroelectric units is considered, taking into account the net head dependence of hydro plants by means of an under-relaxed iterative process. The presented algorithm has been applied to real-scale study cases, obtaining satisfactory results in computational time and optimality. (c) 2006 Elsevier Ltd. All rights reserved.
In this study we model a gene network as a continuous-time switching network. In this model, each gene has a binary state which indicates if the gene expresses or not. We propose a method to control a sequence of expr...
详细信息
In this study we model a gene network as a continuous-time switching network. In this model, each gene has a binary state which indicates if the gene expresses or not. We propose a method to control a sequence of expression patterns in the gene network model by adding another continuous-time switching network. By using propositional calculus, we will show that the control problem can be formulated as a mixed-integer linear programming problem with linear constraints. (c) 2005 Elsevier Ltd. All rights reserved.
Scheduling of resources and tasks has been a key focus of manufacturing-related problems for many years. With increased competition in the global marketplace, manufacturers are faced with reduced profit margins and th...
详细信息
Scheduling of resources and tasks has been a key focus of manufacturing-related problems for many years. With increased competition in the global marketplace, manufacturers are faced with reduced profit margins and the need to increase productivity. One way to meet this need is to implement a flexible manufacturing system (FMS). A FMS is designed such that the efficiency of mass production is achieved while the flexibility of low-volume production is maintained. One type of FMS is the flexible manufacturing cell (FMC), which consists of a group of CNC machines and one material handling device. Scheduling is an important aspect in the overall control of the FMC. This research focuses on production routing and scheduling of jobs within a FMC. The major objective is to develop a methodology that minimizes the manufacturing makespan, which is the maximum completion time of all jobs. Due to the complexity of the FMC routing and scheduling problem, a 0-1 mixed-integer linear programming (MILP) model is formulated for M-machines and N-jobs with alternative routings. Although small instances of the problem can be solved optimally with a commercial optimization software package, a two-stage algorithm is proposed to solve medium-to-large-scale problems more efficiently. This two-stage algorithm utilizes two heuristics to generate an initial feasible sequence and an initial makespan solution during the construction Stage I. Then, during the improvement Stage II, the resulting initial solutions acquired from Stage I are combined with a Tabu Search meta-heuristic procedure. Within the Tabu Search procedure, an efficient pairwise interchange (PI) method and a linearprogramming (LP) subproblem are used to acquire improved *** mathematical model and the proposed two-stage algorithm are demonstrated on several test problems for the makespan performance measure. Although the proposed algorithm does not achieve optimal solutions for every instance, the computational test re
Progression systems synchronize traffic signals to maximize the green band, so that vehicles entering the band at one end of the arterial can pass the other unimpeded. Although traffic control systems optimize other c...
详细信息
Progression systems synchronize traffic signals to maximize the green band, so that vehicles entering the band at one end of the arterial can pass the other unimpeded. Although traffic control systems optimize other criteria, such as travel time and traffic flow, synchronization remains a desirable feature as drivers are averse to frequent stops. This paper gives a short account of progression systems based on bandwidth maximization, including MAXBAND and MULTIBAND , which can be cast as mixed-integer linear programming problems. To this end, the paper offers a brief review of cutting-planes, before developing a Chvátal-Gomory-based cutting-plane algorithm for bandwidth maximization and reporting numerical results. The cutting planes proved to be effective in the first iterations, inducing a substantial reduction on the upper bound for the maximum green band, but became ineffective in the remaining iterations. This paper finalizes with a research agenda to develop fast, effective algorithms for real-time applications.
In this paper we focus on the single-facility capacitated survivable network design problem. We optimize simultaneously the network topology and the link dimensioning in order to route all traffic commodities accordin...
详细信息
In this paper we focus on the single-facility capacitated survivable network design problem. We optimize simultaneously the network topology and the link dimensioning in order to route all traffic commodities according to survivability requirements. The latter are actually expressed in terms of the spare capacity required to address link failures in the context of different rerouting strategies. We present a mixed-integer linear programming model solved by combining several approaches. To tackle the high dimensionality and to separate the continuous and integer variables, we use Benders' decomposition and a cutting-plane approach. Going beyond the proposed method itself, we examine and compare two well-known restoration techniques: local and end-to-end reroutings. Numerous computational results for realistic network instances provide a comparison of these rerouting mechanisms in terms of installed capacities, network density as well as overall costs and CPU time.
The unit commitment problem with tertiary reserve requirements has been broadly studied. Such reserves, when needed, are centrally deployed with relatively slow time constants of the order of minutes. In contrast, the...
详细信息
The unit commitment problem with tertiary reserve requirements has been broadly studied. Such reserves, when needed, are centrally deployed with relatively slow time constants of the order of minutes. In contrast, the scheduling of units offering primary frequency regulation reserve deployable in a decentralized manner within seconds of a contingency has received relatively little attention. In this paper, we formulate and solve a multiperiod unit commitment that simultaneously accounts for both primary and tertiary reserve constraints. What makes this scheduling problem particularly challenging is the characteristic that primary frequency regulation reserves have a common single degree of freedom, namely, the system frequency deviation.
This work presents the development and implementation of. a production scheduling system for an electrical appliance manufacturer. Based on recent advances in optimisation-based scheduling approaches, two different so...
详细信息
This work presents the development and implementation of. a production scheduling system for an electrical appliance manufacturer. Based on recent advances in optimisation-based scheduling approaches, two different software architectures based on two different scheduling formulations, namely the RTN and the STN, are proposed to integrate information available in the different production units and stages with formal algorithmic tools. Optimization results indicate that significant economic benefits can be achieved (e.g. minimization of total operating costs) while ensuring full customer satisfaction as opposed to normal practices followed in the company relying on human expertise. The work indicates that it is possible to solve real-life manufacturing problems using optimization-based approaches but the integration of information in a timely fashion seems to be a major factor in successfully implementing the system and fully realizing its benefits.. (c) 2005 Elsevier Ltd. All rights reserved.
In this paper we describe a mixed-integer linear programming model for a road investment problem. We only focus on the upgrading of gravel roads, hence maintenance and new roads are not included in the model. Furtherm...
详细信息
In this paper we describe a mixed-integer linear programming model for a road investment problem. We only focus on the upgrading of gravel roads, hence maintenance and new roads are not included in the model. Furthermore, we use the model to calculate the optimal level of upgrading over a 10-year planning period, at a forest area owned by a Swedish forest company. We test six different scenarios: that is, three different scenarios on the length of the thawing period in spring for two different objective functions. We conclude that the type of objective function to be used depends on the situation and that the entire network of roads must be optimized simultaneously, to avoid sub-optimizations. Although the forest area includes 440 roads, it is possible to obtain the global optimal solutions from the optimization within a few minutes using commercial software. This is a promising prospect, since the aim is to use the model as part of a future decision support system for strategic planning of road investments. (c) 2004 Elsevier B.V. All rights reserved.
暂无评论