This paper is dedicated to the study of a machining system with standby provisioning, feedback and working vacation policy of the server. The failed machines are permitted by the server in the system according to the FCFS discipline. If no failed machines are found by server, it goes on vacation. But instead of stopping the service altogether, it still operates on a slow service and this process is said to be server's working vacation. When all the standby units are in use, failure of units occur in a degraded mode. The concept of feedback policy is also considered wherein customer (machine) getting service is unsatisfied with the services and wishes to rejoin the system as a feedback customer (machine) or leaves the system. Customer who wishes to provide feedback (unsatisfied customer) can join the queue at the back. The Markov process concept is utilized to present the differential-difference equations of the queuing model. The failure and repair rate of units in machining system is exponentially distributed. Various performance characteristics have been derived. Cost optimization function is evolved for getting the optimal operating conditions. The well-known meta-heuristic technique, particle swarm optimization (PSO) is implemented for obtaining the ideal operating conditions at minimum estimated cost including economic performance. Lastly, numerical results are described related to the studied model through tables and graphs.