Warren H. Hausman

Optimal procurement strategies for online spot markets

Abstract Spot markets have emerged for a broad range of commodities, and companies have started to use them in addition to their traditional, long-term procurement contracts (forward contracts). In comparison to forward contracts, spot markets offer products at essentially negligible lead time, but typically command a higher expected price for this added flexibility while also exhibiting substantial price uncertainty. In our research, we analyze the resulting procurement challenge and quantify the benefits of using spot markets from a supply chain perspective. We develop and solve mathematical models that determine the optimal order quantity to purchase via forward contracts and the optimal quantity to purchase via spot markets. We analyze the most general situation where commodities can be both bought and sold via a spot market and derive closed-form results for this case. We compare the obtained results to the reference scenario of pure contract sourcing and we include results for situations where the use of spot markets is restricted to either buying or selling only. Our approaches can be used by decision makers to determine optimal procurement strategies based on key parameters such as, demand and spot price volatilities, correlation between demand and spot prices, and risk aversion. The results of our analysis demonstrate that significant profit improvements can be achieved if a moderate fraction of the commodity demand is procured via spot markets. The results also show that companies who use spot markets can offer a higher expected service level, but that they might experience a higher variability in profits than companies who do not use spot markets. We illustrate our analytical results with numerical examples throughout the paper.

Scheduling Policies for Automatic Warehousing Systems: Simulation Results

Abstract This paper examines and extends previous analytical work on the scheduling of stacker cranes in automatic warehousing systems. In particular, the following are examined by means of a computer simulation: (1) The performance of the closest-open-location rule compared to random storage assignment; (2) The dynamic behavior of the system under stochastic conditions, using various crane and rack utilization levels; (3) The actual versus predicted reduction in crane travel time due to improved scheduling rules; (4) The effect of imperfect information concerning the length of stay of an incoming pallet on system performance. The results of our experiments provide evidence in support of the proposed analytical models. Most important, the experiments demonstrate the value of previously-proposed scheduling rules in a dynamic, stochastic environment operating with imperfect information.

Joint demand fulfillment probability in a multi-item inventory system with independent order-up-to policies

We consider a multi-item inventory system with dependent item demands represented by a multivariate normal distribution and filled under a First-Come-First-Served rule. Each item is managed independently through a periodic review order-up-to policy while all items have the same review cycle. We obtain the joint demand fulfillment probability within a pre-specified time window. We also study the problem of maximizing the joint demand fulfillment probability and discuss a heuristic approach in which equal safety factors (equal fractiles) are specified for all items. Finally we present numerical results and an application with actual data.

Joint determination of preventive maintenance and safety stocks in an unreliable production environment

In this article we formulate an analytical model of preventive maintenance and safety stock strategies in a production environment subject to random machine breakdowns. Traditionally, preventive maintenance and safety stocks have been independently studied as two separate strategies for coping with machine breakdowns. Our intent is to develop a unified framework so that the two are jointly considered. We illustrate the trade-off between investing in the two options. In addition, we provide optimality conditions under which either one or both strategies should be implemented to minimize the associated cost function. Specifically, cases with deterministic and exponential repair time distributions are analyzed in detail. We include numerical examples to illustrate the determination of optimal strategies for preventive maintenance and safety stocks.

Supply Chain Performance Metrics

Every CEO must always be concerned with the competition. In today’s economy the battlefield is shifting from individual company performance to what we call Supply Chain Performance. Supply Chain Performance refers to the extended supply chain’s activities in meeting end-customer requirements, including product availability, on-time delivery, and all the necessary inventory and capacity in the supply chain to deliver that performance in a responsive manner. Supply Chain Performance crosses company boundaries since it includes basic materials, components, subassemblies and finished products, and distribution through various channels to the end customer. It also crosses traditional functional organization lines such as procurement, manufacturing, distribution, marketing & sales, and research & development.

Multiproduct Production Scheduling for Style Goods with Limited Capacity, Forecast Revisions and Terminal Delivery

A general mathematical model is formulated for the problem of scheduling production quantities for a group of products with seasonal stochastic demand through a common production facility. It is assumed that revised forecasts of total demand over the season for each product become available as the season progresses; delivery is not required until the end of the selling season. Limited production capacity requires that some production take place early in the season, when forecasts are less accurate. At the end of the season, there are overage costs and underage costs representing costs of producing excess quantities and opportunity costs of not producing enough units. Under certain assumptions concerning the data-generating process for forecast revisions, it is possible to formulate the entire problem as a dynamic programming problem; however, the formulation is not computationally feasible if two or more products are considered. Three heuristic approaches to the multi-product problem are presented, and their cost performance is evaluated in some numerical examples. In these particular examples, more frequent reforecasting and rescheduling produces substantial reduction in costs.

Global Logistics Indicators, Supply Chain Metrics, and Bilateral Trade Patterns

Past research into the determinants of international trade highlighted the importance of the basic spatial gravity model augmented by additional variables representing sources of friction. Studies modeled many sources of friction using various proxies, including indices based on expert judgment in some cases. This paper focuses on logistics friction and draws on a data set recently compiled by the World Bank with specific quantitative metrics of logistics performance in terms of time, cost, and variability in time. It finds that the new variables that relate directly to logistics performance have a statistically significant relationship with the level of bilateral trade. It also finds that a single logistics index can capture virtually all of the explanatory power of multiple logistics indicators. The findings should spur public and private agencies that have direct or indirect power over logistics performance to focus attention on reducing sources of friction so as to improve their countrys ability to compete in todays global economy. Moreover, since the logistics metrics are directly related to operational performance, countries can use these metrics to target actions to improve logistics and monitor their progress.

Easy Quantification of Improved Spare Parts Inventory Policies

This paper presents approximate analytical models to quantify the expected improvement in inventory investment when using a system approach to control inventory as opposed to a simpler item approach. A system approach ensures that a demand-weighted average fill rate is achieved at low inventory investment by assigning low fill rates to parts with high costs and high fill rates to parts with low costs. An item approach does not vary fill rates by parts but assigns identical fill rates to all parts. Using single-parameter functional representations of the skewness of unit costs and average demand across all parts in the system, simple approximate analytical expressions for the required inventory investment are derived for both approaches. The accuracy of the approximations is validated using data from a distribution center for computer spare parts. For these data, the solutions obtained by the approximations are very close to the exact values. The results show that inventory investments can be well approximated as a function of only a few cost and demand parameters. These expressions can be used to determine the percentage reduction in inventory investment for a particular target demand-weighted average fill rate when the superior system approach is used instead of the item approach. For increased ease of use, the percentage reduction in inventory when using a system as opposed to an item approach is computed over a range of realistic values for the key parameters of the model and a quadratic expression is fitted to the data. This fitted expression provides rough guidelines for the anticipated improvement with very limited data needed, prior to detailed modeling or implementation.

Multiple failures in a multi-item spares inventory model

Most inventory models dealing with spare parts assume independent part failures and hence one-at-a-time failure occurrences. In practice, for example when an aircraft engine undergoes repair, there is usually more than one part requiring replacement or repair; we denote this situation as ‘multiple failures’. In this paper we represent multiple failures in a continuous, infinite horizon, order-for-order spares replenishment inventory model with cannibalization. We derive exact expressions for the distribution function and the expectation of the number of backlogged jobs; approximations are also provided. The results can be used to optimize spares inventories, to evaluate the effects of correlated failures, and to model the impact of parts commonality.

RFID in mixed-model automotive assembly operations: Process and quality cost savings

This research shows how to characterize the potential operational benefits of RFID in a complex automotive assembly system. Both process savings and quality (rework) savings are included in the presented model and it is demonstrated that process savings and quality cost savings are strongly interdependent. Indeed, a trade-off exists in which the decision maker has to decide whether to allocate time savings afforded by RFID to process time savings, or to allocate those time savings to potential quality cost savings. It is shown how this allocation decision can be solved such that the maximum expected cost savings are achieved. From the structural properties of this optimal solution, it is concluded that an RFID implementation will tend to yield the most benefit in a fast-paced complex assembly environment where: (i) total assembly times are substantial; (ii) tact times are low; (iii) there is little to no worker idle time; and (iv) rework is present and costly. How to evaluate the impact of an RFID implementation in terms of the net present value of cost savings is demonstrated. The model framework is applied in a numerical example to study a situation at a car manufacturer and the main results are reported.