Essays on Time-based Decision Making and the Performance of Lot-sizing Heuristics

Abstract

The productive use of managerial time and the timing of investments present success factors for companies. The popular time management literature recommends investing time now to save time later (e.g., the 5S concept). In finance, early investing is recommended, too, as returns of investments accumulate over time. Managers are often faced with the decision of whether to invest time in a particular task or pay for its completion. In many economic approaches, time is set equal to money, since these are exchangeable via labor wages. However, time is perceived differently than money. In particular, time is perceived to be more ambiguous and less easily fungible than money. Although past investments in a project are irrelevant to deciding whether to invest in the project in the future or not, the occurrence of past investments often influences investment behavior. The tendency to continue with a project when money, effort, or time has already been invested is termed the sunk cost effect and can lead to escalating commitment. The sunk cost effect has been demonstrated in various contexts. However, the results of previous research on the consideration of past investments are inconsistent. In addition to investing time and money, lot sizing represents another topic of operations management and this dissertation. Modern ERP systems allow demand to be forecasted for short time periods. The daily instead of weekly representation impacts demand patterns. Weekly demand may contain only a few periods with zero demand, while periods with zero demand occur more frequently when demand is daily. Close-to-zero demand means that large demands occur at longer intervals, while relatively small demands occur on all other days. In many ERP systems, myopic lot-sizing heuristics like that of Silver & Meal (1973) and Groff (1979) are implemented. However, these heuristics do not perform well for irregular (daily) demand. The dissertation “Essays on Time-based Decision Making and the Performance of Lot-sizing Heuristics” includes 3 research papers. The first two papers analyze how time versus money is invested in a dynamic situation and how time is invested after past time investments occur. In the first paper, we theoretically model and empirically investigate how time versus money is invested in dynamic decision-making situations. In the considered resource allocation problems, early investments are favorable because returns on investments accumulate over time. It is optimal to invest (time/money) and harvest rewards later. However, individuals fail to invest first and harvest later. The central finding is that the timing of investments improves when time investments meet monetary rewards. In these cases, it appears that simple myopic rules do not impose, and cognitive reflection sets in. Paper 2 investigates the effect of the occurrence of past unsuccessful investments. We examine the classical sunk cost situation, where a choice can be made between the sunk cost project and a superior alternative, and the situation where the sunk cost project is the superior project. We analyze whether individuals abandon a project they have unsuccessfully invested time. In the setting considered, without responsibility for past unsuccessful investments, decision makers leave the project with sunk time investments – even if the project is superior. In paper 3, we compare ten lot-sizing heuristics, including well-known simple myopic heuristics like that of Silver & Meal (1973) and Groff (1979), in a rolling horizon environment and find that over all demand patterns and types, the Wagner-Whitin-Look-Beyond algorithm has shown the least mean relative additional costs (0.103 %). The next best heuristic is the combination heuristic K-Gr-zero that consists of the Groff-zero heuristic and an iteratively applied improvement step, in which replenishment points are shifted forward or backward a few periods (mean relative additional costs of 0.87 %).