Integrated frame work for identifying sustainable manufacturing layouts based on big data, machine learning, meta-heuristic and data envelopment analysis

Facility Layout Design (FLP) is an NP-Hard, which is related to the optimization of the arrangement of facilities in a shop floor or manufacturing unit to maximize the performance and minimize the operating cost. An efficient layout must not only consider the profit but also emphasize on other social causes such as energy consumption, pollution, and people's safety. These are referred to as the three pillars of sustainability. This paper identifies the criteria for sustainability in a manufacturing layout and presents an aggregate mathematical formulation for Sustainable Facility Layout Problem (SFLP). The paper proposes a novel 4-stage methodology using Big Data Analytics, Machine Learning, Hybrid Meta-heuristic, Data Envelopment Analysis (DEA), and K-mean clustering for designing an energy-efficient sustainable sub-optimal layout under uncertain (stochastic) demand over multiple periods. The goal of this framework is to 1) uniquely rank and predict an efficient SFLP that overcomes the dimensionality curse associated with handling large data set, 2) use K-mean to identify the criteria that maximally satisfied an efficient sustainable layout and propose an SFLP model that can be customized depending social, political and economic conditions, and 3) evaluate the total energy consumption and CO2 emission for an efficient sustainable facility layout. To show the running methodology a case is presented using hypothetical data from the literature.