A recent revolutionary trend in the retailing industry is that firms are starting to adopt an omni-channel strategy. That is, firms are providing a seamless experience for customers to access multiple channels to execute a purchase. This trend began with the growth of e-commerce, but is now being accelerated by the rise of mobile platforms and devices.

There are many factors that are different in an omni-channel environment versus the traditional single-channel environment. First, customers are increasingly engaging in channel-switching behavior due to differences in price cadences in different channels. Second, inventory and fulfillment are integrated across the retailers' sales channels (e.g., "buy online pickup in store" and "ship from store" programs).

My research in this area is interested in pricing and inventory strategies in an omni-channel retail environment. In our work, we develop a tractable optimization model for joint pricing and inventory allocation in a retailer's multi-location and multi-channel network. This model can easily accommodate any number of business rules that a retailer might have (e.g., sales or price constraints). In another work, we are interested in optimal fulfillment in an omni-channel fulfillment program. Another work we have in this area is about price-matching strategies for an omni-channel retailer with a multitude of omni-channel competitors. We develop a method to calculate a retailer's "value-at-risk" against a competitor and propose price-matching strategies based on this metric. This last work has been filed as a patent by IBM Research.

On November 2013, Super Typhoon Haiyan, the strongest typhoon on record to make landfall, devastated my home country, the Philippines. In its wake, 16 million people lost access to basic necessities. Relief operations were difficult to coordinate due to geographic challenges, lack of infrastructure, limited throughput, and lack of information. Five days after the typhoon, only 20% of the affected population in Tacloban City (the worst hit city) was receiving aid.

I am interested in research involving coordinating emergency operations in a country consisting of a group of islands (an archipelago) such as the Philippines. Many countries that are frequently struck by natural disasters share this feature, making it difficult to coordinate logistics plans for disaster response. We develop tractable models for pre-positioning relief items in preparation for a multi-island disaster under uncertainty on demand and the capacity of the transportation network. We also develop models for last mile distribution in the aftermath of a disaster.

Funded by the Department of Science and Technology (DOST) in the Philippines, a project in collaboration with industry partners in the Philippines we are involved in aims to implement a cloud-based decision support tool based on these models to aid emergency logistics in the Philippines.

[Update 7/29/2015] After several months of development, the first release of the ReliefOps.ph website is live!

[Update 10/26/2015] Our Philippine team held a training session of about 100 regional DSWD (Dept. of Social Welfare and Development) personnel on the ReliefOps.ph platform.

[Update 10/26/2015] Our team just signed a partnership with DSWD, DOST, and IBM Philippines to commit to working together on the ReliefOps.ph platform. Read more about it in this press release. Our work was also mentioned in a recent OpEd piece by Tony La Vina in The Standard.

A gas utility performs maintenance work over its large network of gas pipelines. The utility has a roster of maintenance crews that perform two types of repair work: (i) standard jobs that must be completed before a specified deadline, and (ii) responding to reports of emergency gas leaks that occur randomly throughout the day and could disrupt a crew's schedule and lead to significant overtime. The utility's goal is to perform all standard jobs within the regulatory deadlines, and to address all emergency jobs in a timely manner, and to minimize crew overtime.

In a project with a large multi-state gas utility, we address this major operational challenge. We propose optimization models and algorithms for crew scheduling under random emergencies. Using these models and algorithms, we develop a web-based decision support tool which has been piloted in one of the utility's sites. We estimate a potential 55% reduction in crew overtime using the tool.

A paper we wrote based on this work was awarded First Place in the 2012 Best Paper award given by the INFORMS Service Science Section. An article talking about the project can be found here.