In the manufacturing process of electric rope shovels, an extensive array of components need to be processed. Each component is subject to a distinct sequence of operations, with the number of operations varying by part. Moreover, each of these operations needs to be processed on specific machines within specific processing durations. Therefore, the electric rope shovel production scheduling problem turns out to be challenging for general optimizers, requiring to find the optimal operation sequence, make trade-offs between multiple conflicting objectives, and satisfy a series of strict constraints. To address this production scheduling problem, this paper proposes a neo-cooperation search based evolutionary algorithm. The proposed algorithm suggests a novel encoding scheme to represent a solution (i.e., the sequence of operations of multiple components) with a real decision vector and allocates computational resources to two cooperating populations for global search and local search, respectively. The proposed algorithm can effectively balance between exploration and exploitation, and is shown to outperform state-of-the-art evolutionary algorithms in the experiments.