Profit or Growth? Dynamic Order Allocation in a Hybrid Workforce

Modern digital technology has enabled the emergence of the hybrid workforce in service organizations, where a firm uses on-demand freelancers to augment its traditional labor supply of employees. Freelancers are typically supplied by an electronic platform. How should demand be allocated between employees and freelancers? Under what conditions is the system (comprising the firm and its platform) sustainable in the long run? We investigate these questions in the context of last-mile delivery. We develop a discrete-time, stochastic dynamic program that captures the system’s profit from serving demand and the platform’s growth dynamics. The dynamic model incorporates a service constraint for the platform and a simple version of a stochastic network effect. We find that the answers to our research questions critically depend on two key parameters: the mean and variance of the cross-network effect. We conduct a numerical study with data from a last-mile delivery firm in Vietnam to illustrate our findings. This paper was accepted by Vishal Gaur, operations management.

Individualized Pricing for a Cloud Provider Hosting Interactive Applications

We consider a cloud provider that hosts interactive applications, such as mobile apps and online games. Depending on the traffic of users for an application, the provider commits a subset of its resources (hardware capacity) to serve the application. The provider must choose a dynamic pricing mechanism to indirectly select the applications hosted and maximize revenue. We model the provider’s pricing problem as a large-scale stochastic dynamic program. To approach this problem, we propose a tractable approach to enable decomposing the multidimensional stochastic dynamic program into single-dimensional subproblems. We then extend the proposed framework to define an individualized dynamic pricing mechanism for the cloud provider. We present novel upper bounds on the optimal revenue to evaluate the performance of our pricing mechanism. The computational results show that a contract-based model of selling interactive cloud services achieves significantly greater revenue than the prevalent alternative and that our pricing scheme attains near-optimal revenue.

Dynamic Pricing of Relocating Resources in Large Networks

Motivated by applications in shared vehicle systems, we study dynamic pricing of resources that relocate over a network of locations. Customers with private willingness to pay sequentially request to relocate a resource from one location to another, and a revenue-maximizing service provider sets a price for each request. This problem can be formulated as an infinite-horizon stochastic dynamic program, but it is difficult to solve, as optimal pricing policies may depend on the locations of all resources in the network. We first focus on networks with a hub-and-spoke structure, and we develop a dynamic pricing policy and a performance bound based on a Lagrangian relaxation. This relaxation decomposes the problem over spokes and is thus far easier to solve than the original problem. We analyze the performance of the Lagrangian-based policy and focus on a supply-constrained large network regime in which the number of spokes (n) and the number of resources grow at the same rate. We show that the Lagrangian policy loses no more than O(ln n/n) in performance compared with an optimal policy, thus implying asymptotic optimality as n grows large. We also show that no static policy is asymptotically optimal in the large network regime. Finally, we extend the Lagrangian relaxation to provide upper bounds and policies to general networks with multiple interconnected hubs and spoke-to-spoke connections and to incorporate relocation times. We also examine the performance of the Lagrangian policy and the Lagrangian relaxation bound on some numerical examples, including examples based on data from RideAustin. This paper was accepted by David Simchi-Levi, revenue management and market analytics.

The sequence relay selection strategy based on stochastic dynamic programming

Relay-assisted (RA) network with relay node selection is a kind of effective method to improve the channel capacity and convergence performance. However, most of the existing researches about the relay selection did not consider the statically channel state information and the selection cost. This shortage limited the performance and application of RA network in practical scenarios. In order to overcome this drawback, a sequence relay selection strategy (SRSS) was proposed. And the performance upper bound of SRSS was also analyzed in this paper. Furthermore, in order to make SRSS more practical, a novel threshold determination algorithm based on the stochastic dynamic program (SDP) was given to work with SRSS. Numerical results are also presented to exhibit the performance of SRSS with SDP.

COMPUTING AVERAGE OPTIMAL CONSTRAINED POLICIES IN STOCHASTIC DYNAMIC PROGRAMMING

A stochastic dynamic program incurs two types of cost: a service cost and a quality of service (delay) cost. The objective is to minimize the expected average service cost, subject to a constraint on the average quality of service cost. When the state space S is finite, we show how to compute an optimal policy for the general constrained problem under weak conditions. The development uses a Lagrange multiplier approach and value iteration. When S is denumerably infinite, we give a method for computation of an optimal policy, using a sequence of approximating finite state problems. The method is illustrated with two computational examples.