Parameter selection optimization for parametric cost estimation based on Simulated Annealing

Parametric is the most often used method for Life-Cycle Cost Estimation (LCCE), of which the Parameter Selection Problem (PSP) plays a key role in modeling Cost Estimation Relationships (CERs). It is generally computational infeasible to find out the optimal solution by comparing all the combinations of parameters when the problem is large-sized. Expertise was always counted on in the past. In this paper, we employ the modern meta-heuristic algorithm, i.e., an improved Simulated Annealing algorithm, to solve the problem of large-scale. We also present a mathematic optimization model for the PSP aiming at minimizing the average cost prediction error. A case study is given to show the principle of the proposed model and simulation experiments are carried out to demonstrate effectiveness and efficiency of this algorithm. The results show that this algorithm has a high probability in finding the optimal solution just by searching very small portion of solution space, which is satisfying.