Stochastic optimization for real time service capacity allocation under random service demand

The problem of repeated allocation of limited renewable service resources to distributed service centers is considered here. The objective is to assure a given Quality of Service expressed through percentage of demand which is satisfied during a specified time period. Resource requirements are not fully known at the time when a decision about the service resource distribution is taken. The problem is addressed by formulating a succession of stochastic optimization problems solved at the time of resource allocation. Solutions of these problems are derived by applying duality theory. We pay special attention to the interplay between performance and risk by introducing the concept of a risk budget. Results of numerical experiments confirm the efficiency of the approach.