Move-optimal schedules for parallel machines to minimize total weighted completion time

We study the minimum total weighted completion time problem on identical machines, which is known to be strongly $\mathcal{NP}$-hard. We analyze a simple local search heuristic, moving jobs from one machine to another. The local optima can be shown to be approximately optimal with approximation ratio $1.5$. In case all jobs have equal Smith ratios, the approximation ratio is at most $1.092$