The Integrated Disaster Medical Assistance Team Scheduling and Relief Supply Distribution Problem

In this paper, we study a post-disaster humanitarian logistic problem in which several disaster medical assistance teams or mobile clinics are dispatched to provide medical services to beneficiaries in affected areas. Such... [ view full abstract ]

In this paper, we study a post-disaster humanitarian logistic problem in which several disaster medical assistance teams or mobile clinics are dispatched to provide medical services to beneficiaries in affected areas. Such services often require a certain number of relief supplies that are sourced from various government- and non-government-operated distribution centers. In emergencies, there is not enough time for these assistance teams or mobile clinics to carry sufficient medical supplies such as medicine, vaccines, gauze pads, and syringe needles with them, and therefore the teams are not able to conduct on-site medical services until the arrival of medical supplies. This paper extends the work of Lei et al. published in Annals of Operations Research, 2015, Volume 235, Issue 1, in which the authors assumed that the medical teams could visit another location after completing services at one demand point, but that “the teams’ routes are pre-determined or fixed”. In this paper, however, the traveling routes of assistance teams are not given as inputs, and alternatively they must be determined along with the distribution of relief supplies. Hence, our problem contains a time dependent vehicle routing problem as its sub-problem, making it more complicated. A mixed integer-programming model is first developed to address the issues of equity, efficacy, and efficiency in humanitarian logistics. Then, a two-stage hybrid metaheuristic method is proposed to solve the problem. In the first stage, medical assistance team routes are generated using the Artificial Bee Colony (ABC) algorithm. The fitness function values for feasible solutions are determined through Linear Programming (LP) Relaxation. After the routes are fixed, a Rolling Horizon (RH) approach designed in Lei et al. (2015) is applied to solve the resulting problem. Two other less complex variants of the ABC algorithms in the first stage are also proposed for comparison purposes. Problem instances of various sizes as well as a case study based on the 2016 Kyushu Earthquake in Japan are generated to test our proposed algorithm. Computational results show that the hybrid metaheuristic algorithm is able to find near-optimal solutions in minutes. The performance of the algorithm is also demonstrated to be efficient, especially in emergency situations where quick response is highly desirable.