Asia-Pacific Journal of Operational Research

When decisions are made under uncertainty (DMUU), the decision maker either disposes of an interval of possible profits for each alternative (the interval DMUU) or disposes of a discrete set of payoffs for each decision and then the amount of the profit related to a given alternative depends on the state of nature (the scenario DMUU). Existing methods, used to generate the ranking of decisions and applied to the second problem mentioned, take, to a different extent, into consideration how particular profits assigned to alternatives are ordered in the payoff matrix and what the position of a given outcome is in comparison with other outcomes for the same state of nature. The author proposes and describes several alternative procedures that enable connecting the structure of the payoff matrix with the selected decision. These methods are adjusted to the purpose and the nature of the decision maker. They refer to the Savage’s approach, to the maximin joy criterion, to the normalization technique and to some elements used in expected utility maximization and prospect theory.

With the development of the video-on-demand (VOD) technology, the audiences can decide to skip the advertisements by paying the video platform. This paper examines the optimal pricing decisions for an online video platform by considering the customer’s choice behavior. Based on a two-sided market that is composed of a monopolistic video platform, advertisers and audiences, we show that the platform’s optimal price charging from the advertisers is decreasing in the degree of audiences’ disutility for advertisements, while the optimal price charging from the audiences is convex in the degree of audiences’ disutility for advertisements. More interestingly, we find that the equilibrium number of advertisers on the platform and the video platform’s profit are both convex in the degree of audiences’ disutility for advertisements, which are counter-intuitive. We also extend our model to the case with two duopolistic platforms and find that the optimal number of advertisers on the platform is concave in the degree of audiences’ disutility for advertisements. Numerical experiments are further presented to reveal the managerial insights.

Nowadays, pricing and remanufacturing problems under uncertain markets have gained increasing attention from both industrial and academic fields. In the literature, it is generally assumed that all the channel members are risk-neutral, ignoring the influences of channel members’ risk attitudes in the face of dynamic market. This paper focuses on a pricing problem in a closed-loop supply chain (CLSC) with two competitive risk-sensitive retailers under uncertain environment. The uncertainty is associated with the recycling costs, consumer demands and remanufacturing costs. Due to the dynamic market, supply chain managers may be unable to collect enough historical data to estimate these demands and costs when making pricing and remanufacturing decisions. In such cases, experts’ estimations are usually employed to describe these uncertain parameters. To deal with these human estimations, an uncertainty theory-based model is proposed. Based on the equilibrium results, how the retailers’ risk sensitivity and human estimations (uncertain degrees) affect the prices and profits is analyzed. It is found that both the retailers will get lower profits while the manufacturer will gain more profit when either of the two retailers becomes more risk-averse. We also find that a higher level of uncertainty in the supply chain will induce a higher collecting rate.

This paper studies the joint stock and production capacity rationing polices for a make-to-stock system with two partially substitutable products manufactured at two factories, respectively. Customers are classified into three types with each type further segmented into multiple classes. Type I and Type II customers are selective and accept their favorite products only. Type III accepts either of the two products and hence its demand is regarded as the flexible demand. The management has to decide whether to run or stop production and whether the various classes of demand for each type of customers have to be satisfied from available inventory or not — in which case demand is lost — in order to minimize the firm’s inventory and production costs. We formulate the problem as a make-to-stock queue system and characterize the optimal joint stock and production rationing policies as monotone switching curves. Numerical examples are given to illustrate the optimal policies and demonstrate the value of flexible demand. Moreover, we develop a simple heuristic policy and compare the costs associated with the optimal and heuristic policies for different parameters.

Driven by sensor technologies and Internet of Things, massive real-time data from highly interconnected devices are available, which enables the improvement of decision-making quality. Scheduling of such production systems can be challenging as it must incorporate the latest data and be able to re-plan quickly. In this research, a multi-fidelity model for simultaneous scheduling problem of machines and vehicles at flexible manufacturing system has been proposed. In order to improve the computational efficiency, we extend the framework, called multi-fidelity optimization with ordinal transformation and optimal sampling, with combining with the K-means method. The proposed framework enables the benefits of both fast and inexpensive low-fidelity models with accurate but more expensive high-fidelity models. Results show that this approach can significantly decrease computational cost compared with other algorithms in the literature.

We consider the problem of allocating a set of resources for performing a given collection of jobs. Each of the jobs has specific resource requirement for its execution. Each resource has specific start time, finish time and associated cost per unit usage. The objective is to execute the jobs while incurring minimum cost. The problem is NP-hard. We improve known four-approximation to a three-approximation algorithm. Our algorithm implies better approximation for two other related problems.

Image ranking is to rank images based on some known ranked images. In this paper, we propose an improved linear ordinal distance metric learning approach based on the linear distance metric learning model. By decomposing the distance metric $A$ as $L^{T}L$, the problem can be cast as looking for a linear map between two sets of points in different spaces, meanwhile maintaining some data structures. The ordinal relation of the labels can be maintained via classical multidimensional scaling, a popular tool for dimension reduction in statistics. A least squares fitting term is then introduced to the cost function, which can also maintain the local data structure. The resulting model is an unconstrained problem, and can better fit the data structure. Extensive numerical results demonstrate the improvement of the new approach over the linear distance metric learning model both in speed and ranking performance.

This paper concentrates on improving the convergence properties of the relaxation schemes introduced by Kadrani et al. and Kanzow and Schwartz for mathematical program with equilibrium constraints (MPEC) by weakening the original constraint qualifications. It has been known that MPEC relaxed constant positive-linear dependence (MPEC-RCPLD) is a class of extremely weak constraint qualifications for MPEC, which can be strictly implied by MPEC relaxed constant rank constraint qualification (MPEC-RCRCQ) and MPEC relaxed constant positive-linear dependence (MPEC-rCPLD), of course also by the MPEC constant positive-linear dependence (MPEC-CPLD). We show that any accumulation point of stationary points of these two approximation problems is M-stationarity under the MPEC-RCPLD constraint qualification, and further show that the accumulation point can even be S-stationarity coupled with the asymptotically weak nondegeneracy condition.