Continuous State Dynamic Programming via Nonexpansive Approximation

This paper studies fitted value iteration for continuous state numerical dynamic programming using nonexpansive function approximators. A number of approximation schemes are discussed. The main contribution is to provide error bounds for approximate optimal policies generated by the value iteration algorithm.      

Exact methods and a heuristic for the optimization of an integrated replenishment-storage planning problem

In this paper we study the coordination of different activities in a supply chain issued from a real case. Multiple suppliers send raw materials (RMs) to a distribution center (DC) that delivers them to a unique plant where the storage of the RMs and the finished goods is not possible. Then, the finished goods are directly shipped to multiple customers having just‐in‐time (JIT) demands. Under these hypotheses, we show that the problem can be reduced to multiple suppliers and one DC. Afterwards, we analyze two cases; in the first, we consider an uncapacitated storage at DC, and in the second, we analyze the capacitated storage case. For the first case, we show that the problem is NP‐hard in the ordinary sense using the Knapsack decision problem. We then propose two exact methods: a mixed integer linear program (MILP) and a pseudopolynomial dynamic program. A classical dynamic program and an improved one using the idea of Shaw and Wagelmans are given. With numerical tests we show that the dynamic program gives the optimal solution in reasonable time for quite large instances compared with the MILP. For the second case, the capacity limitation in DC is assumed, which makes the problem solving more challenging. We propose an MILP and a dynamic programming‐based heuristic that provides solutions close to the optimal solution in very short times.     

西门子数字输入中断模块的应用研究

介绍了如何在STEP7中配置西门子数字输入中断模块,如何进行软件中断处理,对中断标志位识别和常见编程处理给出了详细的描述.     

Analogy and duality of texture analysis by harmonics or indicators

According to the classic harmonic approach, an orientation density function (odf)f is expanded into its corresponding Fourier orthogonal series with respect to generalized spherical harmonics, and a pole density function (pdf) into its corresponding Fourier orthogonal series with respect to spherical harmonics. While pdfs are even (antipodally symmetric) functions, odfs are generally not. Therefore, the part of the odf which cannot be determined from normal diffraction pdfs can be mathematically represented as the odd portion of its series expansion. If the odff is given, the even part can be mathematically represented explicitly in terms off itself. Thus, it is possible to render maps ofharmonic orientation ghosts, and to evaluatevariants of mathematical standard odfs resulting in identical pdfs independent of pdf data. However, if only normal diffraction pdfs are known, the data-dependentvariation width of feasible odfs remained unaccessible, and within the harmonic approach a measure of confidence in a solution of the pdf-to-odf inversion problem does not exist.According to any discrete approach, an odff defined on some setG of orientations is expanded into its corresponding Fourier orthogonal series with respect to indicator functions of the elements of a partition ofG, and a pdf defined on the upper (lower) unit hemisphereS ₊ ^{3 3} into its corresponding Fourier orthogonal series with respect to indicator functions of the elements of a partition ofS ₊ ³ . The ambiguity of the pdf-to-odf inversion problem is discussed in terms of column-rank deficiency of the augmented projection matrix. The implication of the harmonic approach to split an odf into auniquely determined and anundetermined part does no longer seem to be reasonable. However, it is possible to numerically determine data-dependent confidence intervals for the Fourier coefficients with respect to the indicator functions which can be immediately interpreted as mean orientation densities within the elements of the partition ofG. Doing so for all Fourier coefficients in the finite series expansion, i.e. for all elements of the partition of the setG, eventually results in the data-dependent variation width of odfs feasible with respect to given pdf data, and to the partitions ofG andS ₊ ³ .Thus it is confirmed that the appearance of orientation ghosts, in particular correlations oftrue andghost orientation components, depends on the representation of an odf. It may be questioned whether in practical applications the implicit assumption of the harmonic method that the even part can be determined uniquely and free of error is generally a reasonable initial condition to develop ghost correction procedures.     

Supervisory control using variable lookahead policies

This paper deals with the efficient on-line calculation of supervisory controls for discrete event systems (DES's) in the framework of limited lookahead control policies (or LLPs) that we introduced in previous papers. In the LLP scheme, the control action after a given trace of events has been executed is calculated on-line on the basis of anN-step ahead projection of the behavior of the DES. To compute these controls, one must calculate after the execution of each event the supremal controllable sublanguage of a finite language with respect to another finite larger language. In our previous work, we showed how the required supremal controllable sublanguage calculation can be performed by using a backward dynamic programming algorithm over the nodes of the tree representation of these two languages. In this paper, we pursue the same approach for the calculation of LLP controls, but instead we adopt a forward calculation procedure over theN-level tree of interest. This forward procedure improves upon previous work by avoiding the explicit consideration of all the nodes of theN-level tree, while still permitting tree-to-tree recursiveness as enabled events are executed by the system. The forward search ends whenever a control decision can be made unambiguously or whenever the boundary of theN-level tree is reached, whichever comes first. This motivates the name Variable Lookahead Policy (or VLP) for this implementation of the LLP supervisory control scheme. This paper presents a general VLP algorithm and studies the properties of several special cases of it. The paper also discusses the implementation of the VLP algorithms and presents computational results regarding the application of these algorithms to a time-varying DES.     

A mathematical model for vehicle routing problem under endogenous uncertainty

In this study, a multistage stochastic programming (SP) model is presented for a variant of single-vehicle routing problem with stochastic demands from a dynamic viewpoint. It is assumed that the actual demand of a customer becomes known only when the customer is visited. This problem falls into the category of SP with endogenous uncertainty and hence, the scenario tree is decision-dependent. Therefore, nonanticipativity of decisions is ensured by conditional constraints making up a large portion of total constraints. Thus, a novel approach is proposed that considerably reduces the problem size without any effect on the solution space. Computational results on some test problems are reported.     

新测试方式下非计算机专业程序设计课程教学探讨

原有教学法在学生实际编程能力上机考试方面存在局限,为适应新的测试方式所带来的教学方法与教学理念的变化,本文结合本校实际,对如何激发与保持学生的学习积极性、实现抽象与模型化的跨越、培养学生的计算思维能力、合理组织实验教学进行探讨,并对实验过程进行监督与综合评价。     

Approximate dynamic programming solutions with a single network adaptive critic for a class of nonlinear systems

Approximate dynamic programming (ADP) formulation implemented with an adaptive critic (AC)-based neural network (NN) structure has evolved as a powerful technique for solving the Hamilton-Jacobi-Bellman (HJB) equations. As interest in ADP and the AC solutions are escalating with time, there is a dire need to consider possible enabling factors for their implementations. A typical AC structure consists of two interacting NNs, which is computationally expensive. In this paper, a new architecture, called the ‘cost-function-based single network adaptive critic (J-SNAC)’ is presented, which eliminates one of the networks in a typical AC structure. This approach is applicable to a wide class of nonlinear systems in engineering. In order to demonstrate the benefits and the control synthesis with the J-SNAC, two problems have been solved with the AC and the J-SNAC approaches. Results are presented, which show savings of about 50% of the computational costs by J-SNAC while having the same accuracy levels of the dual network structure in solving for optimal control. Furthermore, convergence of the J-SNAC iterations, which reduces to a least-squares problem, is discussed; for linear systems, the iterative process is shown to reduce to solving the familiar algebraic Ricatti equation.