University of Southern California

Amy Ward
Helping Companies Do More with Less - It’s all in the Queue
Meet me on LinkedIn
Meet me on Facebook

I am passionate about queueing - or, rather, not queueing. Queues arise in many different application domains, from service systems (like call centers) to computer and communication systems (like web servers) to manufacturing (like assemble-to-order systems). Queues may consist of people or jobs. The common questions that arise in these diverse application domains are: how many resources should we invest in and how can we make real-time decisions to allocate those resources in an efficient manner?

For example, in a call center, the call center manager must decide how many agents to hire, and he must decide how to pair incoming callers, that may have different needs, with agents, that may have different skills. A primary objective is to maintain low waiting times for callers, which occurs when queues are small. The two aforementioned decisions both affect how long the queue of callers waiting to speak to an agent will be. Also, those two decisions are not independent: how incoming callers are routed to agents affects how many agents must be staffed in order to maintain small queues, and how many agents are staffed affects the available pairing choices between agents and callers, which in turn affects the length of the caller queue. When these two decisions are made wisely, the result is low staffing costs and small queues – that is, the call center is able to do more with less resources.

The conclusion is that queues are one way of understanding whether resource investment and allocation decisions are made wisely. In particular, queues illuminate the inherent trade-off between resource investment and cost. Increasing resources (for example, hiring more agents to staff a call center), which is costly, should mean that queues are small. However, if we increase resources and there is still significant queueing, this can indicate that poor resource allocation decisions are being made. Colloquially, we can "do more with less".

Returning to my beginning theme, queues are ubiquitous. This is exactly what sparked my early interest in queueing. I loved that you could learn one methodology, and apply that framework in many different environments, to solve diverse problems. For example, I began in an engineering environment, and more often thought of queues as consisting of jobs. Since arriving at Marshall, and thinking more about business problems, particularly those in the service industry, I now more often think of queues as consisting of people. This shift has been a wonderful way to interact with Marshall students, because they have all queued, and they all share my dislike for waiting in queue! Their feedback and thoughts then stimulates my own research, because they have perspectives that are different from mine, and bring new queueing problems into my awareness.