In this study, we define and formulate the shelter location-allocation problem considering the vulnerability
of the populated locations and their network connectivities with the shelter locations for disaster
management's preparedness and response phase. The shelter provides a safer place and necessary facilities
for victims of a disaster. We formulate the problem as a mixed-integer linear programming (MILP) and stochastic programming model
that selects a set of candidate shelters to the vulnerable populated locations. Thus, the solution presents an
evacuee-allocation plan considering an optimal or the best collection of less vulnerable network
connectivities between the populated areas and the shelter locations. We propose and model the MILP
formulation as a scenario-based stochastic model that assigns the set of candidate locations evaluating
operational, budgetary limitations, and service level expectations for the movement of the desired number
of people from the vulnerable areas. We present a linear relaxation heuristics and compare the heuristics
performance with the scenario-based formulation and two stage stochastic programing formulation solved using CPLEX 12.8 optimization solver for various
problem sizes considering practical application needs and real-life scenarios from Chennai Metropolitan
Development Area. We finally apply and solve the problem using real-life case data obtained during the
major flooding event in and around the Chennai Metropolitan Development Area during 2015 and Nepal earthquake 2015 to present
our model's applicability and emergency response requirements.