This work presents a robust method for the integrated design and operation of batch distillation whereby optimal column sizing, process flexibility and operating policies are obtained simultaneously based on the complex economic trade-offs between capital investment, production revenue and utility costs. The proposed stochastic framework, which utilises a Genetic Algorithm and a penalty function strategy, is found to be successful in obtaining profitable and feasible column designs for many design scenarios including binary and multicomponent mixtures, single duty and multipurpose columns, as well as for regular and complex column configurations. The method can also be used with column models of different complexity. Given a set of design specifications and separation requirements, the optimal number of stages, reboiler duty, reflux profiles, product recoveries, time interval of each distillation tasks, process allocation and number of batches can be obtained. Several design case studies are presented and a comparison of optimal designs for various design scenarios, such as different production time, capital costs, process allocation and mixture characteristics, are discussed. (C) 2003 Elsevier B.V. All rights reserved.