Dynamic Resource Allocation Of Service-oriented Production System With Customer Behavior

Service-oriented manufacturing is a new manufacturing mode that integrates manufacturing and service,which is an important direction of manufacturing transformation and upgrading in the process of industrialization.Service-oriented manufacturing will be beneficial for enterprises to optimize the resource allocation,with the method of guiding the industry transformation from manufacturing to servitization.It can help enterprises improve the supply of products,and also improve the QoS(quality of service)as well as market competitiveness.This paper focuses on the service-oriented manufacturing network system in the uncertain environment,with the consideration of the complex customers behavior.Upstream service-oriented manufacturers provide services for downstream enterprises or customers,and manufacturing service resources can be shared within node enterprises.The time-varying service demand is generated by the customers arrivals,whom are not only sensitive to the waiting time,but also sensitive to the price of service providers.Customers will randomly give up queueing when waiting for service,after that there will be a random retry behavior.And there will be a random re-purchase(feedback)behavior after completing the service.The complex behavior of customers is considered for the service-oriented manufacturing model.On the system premise of satisfying the service level,we propose a dynamic resource allocation strategy based on queuing theory,which can ensure the reasonable utilization of manufacturing resources and the stabilization of customer abandonment.Measure quality of service with system performance.Firstly,considering the dynamic resource allocation of each manufacturing node for a multi-level SMN(service-oriented manufacturing network),we propose the allocation strategy of SMN under low-QoS target based on(delayed infinite server network – offered load,DISN-OL)approximation and allocation strategy of SMN under high-QoS target based on(delayed infinite server network – modified offered load,DISN-MOL)approximation.Then considering the dynamic resource allocation of a single node for a PSM(pricing service-oriented manufacturing)system,we propose the allocation strategy of PSM under low-QoS target based on P-DIS-OL(pricing delayed infinite server offered load)approximation and allocation strategy of PSM under high-QoS target based on P-DIS-MOL(pricing delayed infinite server modified offered load)approximation.Lastly,considering the dynamic resource allocation of a single node in the CBSM(service-oriented manufacturing with complex behavior)system.For one time retrial loop and one time feedback loop model,we propose the allocation strategy of CBSM under low-QoS target based on DIS-OL approximation and allocation strategy of CBSM under high-QoS target based on DIS-MOL approximation.and that For non-constraint loop scenario,we propose the simulation-based ISA(iterative staffing algorithm)approximate algorithm.Considering the various service-oriented manufacturing system with different QoS target,we propose the approximation algorithms and conduct the numerical simulation to dynamically allocate service resources for each manufacturer.The simulation results of numerical examples show that the proposed strategies can effectively control and stabilize the target system performance.This paper provides a dynamic resources allocation strategy for service-oriented manufacturing system,whose approximation is highly consistent with the simulation results and can be extended for the more complex system.