Multi-objective evolutionary algorithms are generally good: Maximizing monotone submodular functions over sequences

Evolutionary algorithms (EAs) are general-purpose optimization algorithms, inspired by natural evolution. Recent theoretical studies have shown that EAs can achieve good approximation guarantees for solving the problem classes of submodular optimization, which a wide range applications, such as maximum coverage, sparse regression, influence maximization, document summarization and sensor placement, just to name few. Though they provided some explanation nature EAs, considered objective functions defined only over sets or multisets. To complement this line research, paper class maximizing monotone sequences, where function depends on order items. We prove each kind previously studied i.e., prefix functions, weakly strongly DAG simple multi-objective EA, GSEMO, always reach improve best known guarantee after running polynomial time in expectation. Note these best-known be obtained different greedy-style before. Empirical various e.g., accomplishing tasks, information gain, search-and-tracking recommender systems, show excellent performance GSEMO.