An Extended Agent Based Model for Service Delivery Optimization

Service delivery optimization has an important impact on organizational profitability, where changes in allocation of resources (e.g. humans, equipment and materials) to services increases profit. Simulation and optimization techniques generally suffer from three main drawbacks; firstly, the limited knowledge and skill of researchers in modeling social complexities. Secondly, having assumed that a fairly realistic model of the problem is simulated, finding optimal solutions requires an exhaustive search that is almost impossible in problems with a large search space. Thirdly, mathematical optimization techniques often require the acquisition of knowledge in a central unit, which is problematic e.g. for privacy reasons. This article introduces a new technique, which combines Agent Based Modeling (ABM) and Distribution Constraint Optimization (DCOP) to overcome these difficulties. Our empirical results present a successful model for finding optimized resourced allocation settings in comparison with two different ABM simulated models on a sample of a real-life service delivery problem.