A stochastic programming approach for planning horizons of infinite horizon capacity planning problems

Planning horizon is a key issue in production planning. Different from previous approaches based on Markov Decision Processes, we study the planning horizon of capacity planning problems within the framework of stochastic programming. We first consider an infinite horizon stochastic capacity planning model involving a single resource, linear cost structure, and discrete distributions for general stochastic cost and demand data (non-Markovian and non-stationary). We give sufficient conditions for the existence of an optimal solution. Furthermore, we study the monotonicity property of the finite horizon approximation of the original problem. We show that, the optimal objective value and solution of the finite horizon approximation problem will converge to the optimal objective value and solution of the infinite horizon problem, when the time horizon goes to infinity. These convergence results, together with the integrality of decision variables, imply the existence of a planning horizon. We also develop a useful formula to calculate an upper bound on the planning horizon. Then by decomposition, we show the existence of a planning horizon for a class of very general stochastic capacity planning problems, which have complicated decision structure.     

Finite horizon approximations of infinite horizon linear programs

This paper describes the class of infinite horizon linear programs that have finite optimal values. A sequence of finite horizon (T period) problems is shown to approximate the infinite horizon problems in the following sense: the optimal values of theT period problems converge monotonically to the optimal value of the infinite problem and the limit of any convergent subsequence of initialT period optimal decisions is an optimal decision for the infinite horizon problem.     

Approximate solutions to infinite dimensional LQ problems over infinite time horizon

This paper is addressed to develop an approximate method to solve a class of infinite dimensional LQ optimal regulator problems over infinite time horizon. Our algorithm is based on a construction of approximate solutions which solve some finite dimensional LQ optimal regulator problems over finite time horizon, and it is shown that these approximate solutions converge strongly to the desired solution in the double limit sense.     

Approximation solution and suboptimality for discounted semi-markov decision problems with countable state space

In this article we consider the approximate solution for semi-Markov decision problems with infinite horizon, countable state space, discounted cost function and finite action space. We present converging sequences of lower and upper bounds for the value function and, moreover, we derive a method for exclusion of suboptimal actions.     

需求不确定的多周期的容量决策问题

研究每个周期的需求随机增加的情形下的容量扩充问题,建立起切合实际的有限周期随机动态规划模型及在期现值准则下的无限周期随机动态规划模型,进而探索生产单一产品的公司在面对随机增加的市场需求时,风险中立的管理者该如何扩充其生产容量,才能使得其公司在折扣意义下的总期望利润最大.研究无限阶段的容量扩充问题,得出某种约束条件下的优化策略解,给公司管理者提供了其长期可持续发展的优化策略和依据.     

Sequencing with uncertain numerical data for makespan minimisation

We consider a scheduling problem with the objective of minimising the makespan under uncertain numerical input data (for example, the processing time of an operation, the job release time and due date) and fixed structural input data (for example the precedence and capacity constraints). We assume that at (before) the scheduling stage the structural input data are known and fixed but all we know about the numerical input data are their upper and lower bounds, where the uncertain numerical data become realised at the control stage as the scheduled process evolves. After improving the mixed graph model, we present an approach for dealing with our scheduling problem under uncertain numerical data based on a stability analysis of an optimal makespan schedule. In particular, we investigate the candidate set of the critical paths in a circuit-free digraph, characterise a minimal set of the optimal schedules, and develop an optimal and a heuristic algorithm. We also report computational results for randomly generated as well as well-known test problems.     

Maximum-throughput dynamic network flows

This paper presents and solves the maximum throughput dynamic network flow problem, an infinite horizon integer programming problem which involves network flows evolving over time. The model is a finite network in which the flow on each arc not only has an associated upper and lower bound but also an associated transit time. Flow is to be sent through the network in each period so as to satisfy the upper and lower bounds and conservation of flow at each node from some fixed period on. The objective is to maximize the throughput, the net flow circulating in the network in a given period, and this throughput is shown to be the same in each period. We demonstrate that among those flows with maximum throughput there is a flow which repeats every period. Moreover, a duality result shows the maximum throughput equals the minimum capacity of an appropriately defined cut. A special case of the maximum dynamic network flow problem is the problem of minimizing the number of vehicles to meet a fixed periodic schedule. Moreover, the elegantsolution derived by Ford and Fulkerson for the finite horizon maximum dynamic flow problem may be viewed as a special case of the infinite horizon maximum dynamic flow problem and the optimality of solutions which repeat every period.     

Step-guided strategies for control problems with uncertain dynamics

We study here a rather general class of nonlinear control systems with uncertain dynamics involving perturbating variables, without assuming any statistical information on them. Contrary to majority of publications on uncertain control problems, we study here the value function and seek for optimal (-optimal) control policies. Our four theorems, covering all possible situations that may happen in the context of relationships between the influence of the control parameters on the system versus the perturbating parameters, give both upper and lower bounds for the value function in terms of solutions of simplified unperturbed control problems. They also provide constructive procedures for finding strategies that, under additional assumptions, appear to be -optimal control policies in the class of so-called step guided strategies introduced by us in a very recent paper.     

On upper bounds of sequential stochastic production planning problems

In this paper we consider a class of problems that determine production, inventory and work force levels for a firm in order to meet fluctuating demand requirements. A production planning problem arises because of the need to match, at the firm level, supply and demand efficiently. In practice, the two common approaches to counter demand uncertainties are (i) carrying a constant safety stock from period to period, and (ii) planning with a rolling horizon. Under the rolling horizon (or sequential) strategy the planning model is repeatedly solved, usually at the end of every time period, as new information becomes available and is used to update the model parameters. The costs associated with a rolling horizon strategy are hard to compute a priori because the solution of the model in any intermediate time period depends on the actual demands of the previous periods.In this paper we derive two a priori upper bounds on the costs for a class of production planning problems under the rolling horizon strategy. These upper bounds are derived by establishing correspondences between the rolling horizon problems and related deterministic programs. One of the upper bounds is obtained through Lagrangian relaxation of the service level constraint. We propose refinements to the non-Lagrangian bounds and present limited computational results. Extensions of the main results to the multiple item problems are also discussed. The results of this paper are intended to support production managers in estimating the production costs and value of demand information under a rolling horizon strategy.     

Bounds on the value of information in uncertain decision problems II†

In most stochastic decision problems one has the opportunity to collect information that would partially or totally eliminate the inherent uncertainty. One wishes to compare the cost and value of such information in terms of the decision maker's preferences to determine an optimal information gathering plan. The calculation of the value of information generally involves oneor more stochastic recourse problems as well as one or more expected value distribution problems. The difficulty and costs of obtaining solutions to these problems has led to a focus on the development of upper and lower bounds on the various subproblems that yield bounds on the value of information. In this paper we discuss published and new bounds for static problems with linear and concave preference functions for partial and perfect information. We also provide numerical examples utilizing simple production and investment problems that illustrate the calculations involved in the computation of the various bounds and provide a setting for a comparison of the bounds that yields some tentative guidelines for their use. The bounds compared are the Jensen's Inequality bound,the Conditional Jensen's Inequality bound and the Generalized Jensen and Edmundson-Madansky bounds.     

Rigorous solution of linear programming problems with uncertain data

This note gives a synopsis of new methods for solving linear systems and linear programming problems with uncertain data. All input data can vary between given lower and upper bounds. The methods calculate very sharp and guaranteed error bounds for the solution set of those problems and permit a rigorous sensitivity analysis. For problems with exact input data in general the calculated bounds differ only in the last bit in the mantissa, i.e. they are of maximum accuracy. This is a written account of an invited lecture delivered by the second author on occasion of the 14. Symposium über Operations Research, Ulm, 6.–8.9.1989.     

Optimal greedy policies for stochastic control models

We introduce the notion of a greedy policy for general stochastic control models. Sufficient conditions for the optimality of the greedy policy for finite and infinite horizon are given. Moreover, we derive error bounds if the greedy policy is not optimal. The main results are illustrated by Bayesian information models, discounted Bayesian search problems, stochastic scheduling problems, single-server queueing networks and deterministic dynamic programs.     

Optimal Policies for Production/Inventory Systems with Finite Capacity and Markov-Modulated Demand and Supply Processes

In many production/inventory systems, not only is the production/inventory capacity finite, but the systems are also subject to random production yields that are influenced by factors such as breakdowns, repairs, maintenance, learning, and the introduction of new technologies. In this paper, we consider a single-item, single-location, periodic-review model with finite capacity and Markov modulated demand and supply processes. When demand and supply processes are driven by two independent, discrete-time, finite-state, time-homogeneous Markov chains, we show that a modified, state-dependent, inflated base-stock policy is optimal for both the finite and infinite horizon planning problems. We also show that the finite-horizon solution converges to the infinite-horizon solution.     

A Continuous Time Financial Market with a Poisson Process as Transaction Timer

We study the influence of transactions on investors' portfolio values under the assumption that the investors' transaction times are determined by Poisson point processes, whose intensity measures can naturally be interpreted as transaction frequencies. We give lower and upper bounds on the expectations of portfolio values in terms of transaction intensities, and prove that there exist a sequence of portfolio fractions and a transaction frequency which maximize the expectation of the portfolio value for a finite horizon. We also give bounds on transaction frequencies for preventing the investors from losing money. Then the optimal transaction strategies for finite and infinite time horizons and the asymptotic effects of making transactions are discussed based on the concept of a benefit function of transactions. Finally, we investigate the influence of transactions on financial markets, with the market mean rates of return and volatilities being connected with the transaction frequency. We show that the market becomes unprofitable in a finite time if an overwhelming amount of transactions is involved and the market is suitable for some limited transactions when its trade capacity does not increase beyond any limit at a relatively high speed. Our models and simulations illustrate how the investors' collective action affects the financial market.     

Planning horizons for capacity expansion

This paper develops procedures for determining the appropriate length of the forecast horizon in the context of a capacity expansion problem. An infinite horizon framework would, in general, render the problem intractable. An arbitrary choice of the horizon length would no longer guarantee the optimal solution. In the context of a rolling horizon procedure, it is only the initial decision that is relevant. The approach, therefore, is to solve longer and longer horizon problems until the optimal initial decision stabilizes. A finite set of capacity additions is available. The cost trade-off is between economies of scale and discounting effects.     

Risk-sensitive control of continuous time Markov chains

We study risk-sensitive control of continuous time Markov chains taking values in discrete state space. We study both finite and infinite horizon problems. In the finite horizon problem we characterize the value function via Hamilton Jacobi Bellman equation and obtain an optimal Markov control. We do the same for infinite horizon discounted cost case. In the infinite horizon average cost case we establish the existence of an optimal stationary control under certain Lyapunov condition. We also develop a policy iteration algorithm for finding an optimal control.     

Optimal control problem of the uncertain second‐order circuit based on first hitting criteria

First hitting criteria of a system are to initially achieve some performance indeces of the target state set. This paper primarily investigates the optimal control problem of the uncertain second‐order circuit based on first hitting criteria. First, considering time efficiency and different from the ordinary expected utility criterion over an infinite time horizon, two first hitting criteria which are reliability index and reliable time criteria are innovatively proposed. Second, assuming the circuit output voltage as an uncertain variable when the historical data is lacking, we better model the real circuit system with the uncertain second‐order differential equation which is essentially the uncertain fractional‐order differential equation. Then, based on the first hitting time theorem of the uncertain fractional‐order differential equation, the distribution function of the first hitting time under the second‐order circuit system is proposed and the uncertain second‐order circuit optimal control model (reliability index and reliable time‐based model) is transformed into corresponding crisp optimal problem. Lastly, analytic expressions of the optimal control for the reliability index model are obtained. Meanwhile, sufficient condition and guidance for parameters for the optimal solution of the reliable time‐based model are derived, and corresponding numerical examples are also given to demonstrate the fluctuation of our optimal solution for different parameters.     

Controlled jump processes

Finite and infinite planning horizon Markov decision problems are formulated for a class of jump processes with general state and action spaces and controls which are measurable functions on the time axis taking values in an appropriate metrizable vector space. For the finite horizon problem, the maximum expected reward is the unique solution, which exists, of a certain differential equation and is a strongly continuous function in the space of upper semi-continuous functions. A necessary and sufficient condition is provided for an admissible control to be optimal, and a sufficient condition is provided for the existence of a measurable optimal policy. For the infinite horizon problem, the maximum expected total reward is the fixed point of a certain operator on the space of upper semi-continuous functions. A stationary policy is optimal over all measurable policies in the transient and discounted cases as well as, with certain added conditions, in the positive and negative cases.     

Analytic solutions for infinite horizon stochastic optimal control problems via finite horizon approximation: A practical guide

We show how infinite horizon stochastic optimal control problems can be solved via studying their finite horizon approximations. This often leads to analytical solutions for the infinite horizon problem by studying phase diagrams, even in cases where the complexity of the finite horizon case does not permit analytic solutions. Our approach can be applied to many problems in dynamic economics.     

Numerical methods for fourth‐order nonlinear elliptic boundary value problems

The aim of this article is to present several computational algorithms for numerical solutions of a nonlinear finite difference system that represents a finite difference approximation of a class of fourth‐order elliptic boundary value problems. The numerical algorithms are based on the method of upper and lower solutions and its associated monotone iterations. Three linear monotone iterative schemes are given, and each iterative scheme yields two sequences, which converge monotonically from above and below, respectively, to a maximal solution and a minimal solution of the finite difference system. This monotone convergence property leads to upper and lower bounds of the solution in each iteration as well as an existence‐comparison theorem for the finite difference system. Sufficient conditions for the uniqueness of the solution and some techniques for the construction of upper and lower solutions are obtained, and numerical results for a two‐point boundary‐value problem with known analytical solution are given. © 2001 John Wiley & Sons, Inc. Numer Methods Partial Differential Eq 17:347–368, 2001