“…case (where the X i 's are independent and 12 see e.g. Tucci [366], Teh et al [358], Aghamohammadi & Crutchfield [6], Herrero-Collantes & Garcia-Escartin [159], Balygin et al [31], Dang et al [103], Gong et al [138], Chandrasekaran et al [77], Drahi et al [117], Kollmitzer et al [194], Liu et al [228], Fischer & Gauthier [126], Kim et al [188], Stoller & Campbell [342] 13 a detailed discussion and comparisons are beyond the scope of this paper, given its current length 14 see e.g. Schrijver [324], Bertsimas & Weismantel [45], Chen et al [83], Onn [279], Korte & Vygen [195], Wolsey [393] for comprehensive books on discrete, integer and combinatorial programming and their vast applications 15 notice that P emp n a probability measure on the data space Y, which is random due to its dependence on the X i 's have common distribution P), ergodic Markov chains on Y with stationary distribution P, more globally autoregressive chains with stationary measure P, etc.…”