Ramesh Bollapragada

Managing Inventory and Supply Performance in Assembly Systems with Random Supply Capacity and Demand

We consider stock positioning in a pure assembly system controlled using installation base-stock policies. When component suppliers have random capacity and end-product demand is uncertain, we characterize the systems inventory dynamics. We show that components and the end product play convex complementary roles in providing customer service. We propose a decomposition approach that uses an internal service level to independently determine near-optimal stock levels for each component. Compared with the optimal, the average error of the decomposition approach is 0.66% across the tested instances. Compared with current practice, this approach has the potential to reduce the safety-stock cost by as much as 30%. Our computational analysis on two-echelon systems also illustrates several managerial insights: We observe that the cost reduction from improving supply performance is high when demand variability or the number of components or target customer service is high, or when the end product is more expensive relative to components. On average, (i) reducing the lead time of the more expensive component yielded higher benefit than reducing the lead time for the less expensive component, and (ii) the benefit of improving one of the supply parameters (service level or lead time) was higher when the value of the other parameter was already more favorable (lower lead time or higher service level, respectively).Finally, we analytically show how a multi-echelon pure assembly system may be converted into an equivalent two-echelon assembly system to which all our results apply.

Replenishment planning in discrete-time, capacitated, non-stationary, stochastic inventory systems

In this paper, we examine a single-product, discrete-time, non-stationary, inventory replenishment problem with both supply and demand uncertainty, capacity limits on replenishment quantities, and service level requirements. A scenario-based stochastic program for the static, finite-horizon problem is presented to determine replenishment orders over the horizon. We propose a heuristic that is based on the first two moments of the random variables and a normal approximation, whose solution is compared with the optimal from a simulation-based optimization method. Computational experiments show that the heuristic performs very well (within 0.25% of optimal, on average) even when the uncertainty is non-normal or when there are periods without any supply. We also present insights obtained from sensitivity analyses on the effects of supply parameters, shortage penalty costs, capacity limits, and demand variance. A rolling-horizon implementation is illustrated.

Budget-Constrained, Capacitated Hub Location to Maximize Expected Demand Coverage in Fixed-Wireless Telecommunication Networks

This paper presents a quantitative model for telecommunication network installation by companies in the broadband-access business, specialized to the fixed-wireless case. Under stochastic demand modeled using scenarios, we maximize the expected demand coverage subject to a budget constraint on hub installation, and technological constraints on demand coverage by installed hubs. There are multiple hub types, differing in costs and capacities. We present a practical greedy heuristic based on the budgeted maximum-coverage problem and analyze its worst-case performance. For special cases with a single hub type or a single demand scenario, we show that a guarantee of 1 - 1/e or 63.2% applies to our greedy heuristic. For the general case we develop a data-dependent performance guarantee. Through computational experiments, we show that the greedy heuristics empirical performance is, on average, within 2% of the optimal expected demand coverage.

An empirical study of policies to integrate reactive scheduling and control in just-in-time job shop environments

In this paper, we compare the performance of policies for integrating reactive scheduling and control that differ in the way they interpret and dynamically reoptimize schedules in the face of contingencies. We conduct our analysis in the context of just-in-time job shop environments ( job shop problems with an objective of minimizing the sum of tardiness and inventory costs), subject to machine failures. We empirically evaluate the tradeoffs in schedule quality and computational time of different scheduling policies under different load conditions and different levels of uncertainty. Our results show that reactive procedures that selectively reoptimize a subset of the scheduling problems are capable of producing high-quality solutions in a fraction of the time required to generate brand new schedules.

Role of Random Capacity Risk and the Retailer in Decentralized Supply Chains with Competing Suppliers

This research considers a supply chain under the following conditions: (i) two heterogeneous suppliers are in competition, (ii) supply capacity is random and pricing is endogenous, (iii) consumer demand, with and without an intermediate retailer, is price dependent. Specifically, we examine how uncertainty in supply capacity affects optimal ordering and pricing decisions, supplier and retailer profits, and the incentives to reduce such uncertainty. When two suppliers sell through a monopolistic retailer, supply uncertainty not only affects the retailers diversification strategy for replenishment, but also changes the suppliers� wholesale price competition and the incentive for reducing capacity uncertainty. In this dual-sourcing model, we show that the benefit of reducing capacity uncertainty depends on the cost heterogeneity between the suppliers. In addition, we show that a supplier does not necessarily benefit from capacity variability reduction. We contrast this incentive misalignment with findings from the single-supplier case and a supplier-duopoly case where both suppliers sell directly to market without the monopolistic retailer. In the latter single-supplier and duopoly cases, we prove that the unreliable supplier always benefits from reducing capacity variability. These results highlight the role of the retailers diversification strategy in distorting a suppliers incentive for reducing capacity uncertainty under supplier price competition.

A two-phase greedy algorithm to locate and allocate hubs for fixed-wireless broadband access

We study a two-phase, budget-constrained, network-planning problem with multiple hub types and demand scenarios. In each phase, we install (or move) capacitated hubs on selected buildings. We allocate hubs to realized demands, under technological constraints. We present a greedy algorithm to maximize expected demand covered and computationally study its performance.

Component procurement and end product assembly in an uncertain supply and demand environment

In this paper, we examine a multi-product, multi-component, procurement and assembly problem with both supply and demand uncertainties. We explicitly model the uncertainty using a stochastic program that facilitates procurement and assembly decisions. We present a stochastic linear programming model of the problem which we solve using its deterministic equivalent with a finite number of scenarios. We show that the systems performance, i.e. total cost, depends on the order lead time, L, and the assembly lead time Q, only through their sum . We illustrate the impact of on system performance with numerical experiments. Also, we illustrate that, in some cases, increasing supplier capacity can worsen system performance. In addition, we identify the key cost drivers that need attention from managers in the manufacturing industry, when there is limited knowledge of future demand and component availability.

Constraint-Based Local Search for Inventory Control Under Stochastic Demand and Lead Time

In this paper, we address the general multiperiod production/inventory problem with nonstationary stochastic demand and supplier lead time under service-level constraints. A replenishment cycle policy is modeled. We propose two hybrid algorithms that blend constraint programming and local search for computing near-optimal policy parameters. Both algorithms rely on a coordinate descent local search strategy; what differs is the way this strategy interacts with the constraint programming solver. These two heuristics are first, compared for small instances against an existing optimal solution method. Second, they are tested and compared with each other in terms of solution quality and run time on a set of larger instances that are intractable for the exact approach. Our numerical experiments show the effectiveness of our methods.

Discrete-Time, Economic Lot Scheduling Problem on Multiple, Non-Identical Production Lines

This paper deals with a general discrete time dynamic demand model to solve real time resource allocation and lot-sizing problems in a multimachine environment. In particular, the problem of apportioning item production to distinct manufacturing lines with different costs (production, setup and inventory) and capabilities is considered. Three models with different cost definitions are introduced, and a set of algorithms able to handle all the problems are developed. The computational results show that the best of the developed approaches is able to handle problems with up to 10000 binary variables outperforming general-purpose solvers and other randomized approaches. The gap between lower and upper bound procedures is within 1.0% after about 500 seconds of CPU time on a 2.66Â Ghz Intel Core2 PC.

Network Planning of Broadband Wireless Networks

To deploy broadband networks, service providers, such as competing local exchange carriers, need robust plans for providing various types, amounts, and locations of services at competitive prices. Broadband networks generally consist of an access component (wireless access), a concentration component (a wireless aggregation point or hub), a service routing or distribution component (a central office or metro switch), and various combined or separate distribution components (a long-haul backbone data or voice network). Because access, aggregation, and routing or distribution vary greatly in requirements, we developed a method and platform for planning the components of fixed-wireless-broadband (FWB) systems for local loop access. We have helped various service providers to analyze and design many networking scenarios using our methods. The service providers have used these scenarios and their predicted financial outcomes to plan FWB access networks tailored to meet their marketing and financial goals. By implementing our method, one service provider has improved its planning process, achieved a competitive advantage in its markets, and increased its annual service revenues by tens of millions of dollars.