Fuzzy robust optimization for handling feed stream and model parameter uncertainties during comminution process

With respect to the uncertainty in the model parameters, the Pareto optimal solutions for the deterministic equivalent of the original uncertain optimization problem have four objective functions: maximization of expected values of throughput and mid-size fractions and minimization of standard deviation of throughput and mid-size fraction as shown in the figure for different types of fuzzy functions. “Type a” and “Type b” correspond to the fuzzy functions which are generated by allowing a deviation of 25% on the lower and upper side of the deterministic value, respectively. “Type b” uncertain values are showing better Pareto optimal solutions compared to “Type a” uncertain values while considering uncertainty in model parameters. The trend is opposite in case of operational parameters. Considering uncertain parameters one at a time, the magnitude of shift in Pareto front helps in identifying the most sensitive uncertain parameter. 139