Heterogeneous extractive batch distillation of the chloroform-methanol minimum-boiling-temperature azeotropic mixture is studied with water as a heterogeneous entrainer. The continuous feeding of water allows recovery of 99 mol 96 chloroform after condensation of the saddle binary heteroazeotrope water-chloroform. Unlike a homogeneous process, the reflux composition is different from the heteroazeotrope water-chloroform in the vapor overhead. Besides, the distillate recovery is improved by refluxing a portion alpha of the chloroform-rich distillate phase along with the water-rich entrainer phase. A genetic algorithm is coupled to a constant molar overflow model to study seven operation policies. Optimization parameters are the entrainer flow rate F-E/V, the portion alpha(T3) of the distillate-rich phase refluxed to the column during chloroform distillation, and the reflux policy RT5 during methanol distillation. The optimization maximizes a profit function, penalized with recovery yield constraints, whereas purity targets are used as task ending events. All optimized solutions achieve higher than 90 mol % recovery yields and 99 mol % purity for both products chloroform and methanol. The results are confirmed by rigorous simulation showing the good performance of coupling a simplified model and a genetic algorithm as a first approach. The two-piece-wise parameter value operation policies for all three parameters increase profit by 41.8% and reduce total time by 43.8% compared to the single value parameter policy. The parameter influence study ranks R-T5 first, alpha(T3) second, and F-E/V last. Keeping F-E/V constant and using the two-piece-wise alpha T-3 and R-T5 value operation is recommended to increase profit.