Simulating job-search models using simulating annealing

Paper deals with a simulation of job-search models using simulated annealing (SA). It is standard to test Burdett-Mortensen on-the-job search model which gives a unique solution. Unfortunately, its later generalizations especially incorporation of heterogeneous agents has overcome this unique equilibrium point for the generalized model could produce multiple equilibriums as well. In such cases many statistical methods using conventional numerical optimization algorithms could fail to find a global optimum point. Simulated annealing, instead, is very robust method capable of finding global optimum point also on very difficult non-linear functions. The method could be applied in discrete or continuous global optimization problems. Its recent modifications, i.e. Adaptive Simulated Annealing, have shortened the convergence time to converge in real time which has made the method widely applicable. Paper doesn’t give additional improvements of the simulated annealing method but applies it to the job-search models