Abstract: Decisions to support preparedness activities for disastermanagement are challenging due to the uncertainties of parameters, the balance preparedness and risk, so it is a hot topic. In this paper multiple relief resources location-routing problem is addressed to determine optimal deployment of supply facilities for multiple relief resources, transport distribution route. However, traditional methods addressing this problem mainly focus on stochastic optimization by assuming probability distribution to measure the uncertainty, there are some drawbacks. Multiple relief resources cost uncertainty is considered, introducing two types of uncertainty sets, i.e. box and ellipsoid, to capture the uncertain cost of multiple relief resources, and multiple relief resources robust location-routing models are proposed respectively, which are converted into the deterministic robust equivalent models, and can be solved by hybrid programming algorithm coded in GAMS and CPLEX. Finally, the 19 cities in the north and west of Sichuan Province are chosen to conduct the numerical study. Results show that the proposed robust models is feasible and effective, and compared to the robust model based on box uncertainty set, under the same parameter setting, robust model based on ellipsoid uncertainty set usually is more conservative, and leads to a higher total cost. Decision-makers, according to their risk aversion and conservativeness, choose an appropriate value for the uncertain level parameters Γ/Ω to get the optimal solution, and provide decision support to the department of Emergency Disaster Relief.

Key words: relief resource, location-routing, cost uncertainty, robust optimization, uncertainty set