A divide and conquer approach for web services location allocation problem
Harshal Tupsamudre, Saket Saurabh, Arun Ramamurthy, Mangesh Gharote, Sachin Lodha · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2021
The appropriate choice of locations for the deployment of web services is of significant importance. The placement of a web service closer to user centers minimizes the response time, however deployment cost may increase. The placement becomes more challenging when multiple web services are involved. In this paper, we address the problem of placing multiple web services with the aim of simultaneously minimizing conflicting objectives of total deployment cost and network latency. We solve the location allocation problem for each web service independently and combine the resulting solutions using a novel merge algorithm. We demonstrate through extensive experiments and simulations that the proposed approach is not only computationally efficient but also produces good quality solutions. Further, the proposed merge algorithm is generic and could be easily adapted to tackle any bi-objective optimization problem that can be decomposed into non-overlapping sub-problems.