This　paper　investigates　a　real　life　bi-objective　hybrid　flow　shop　scheduling　problem　in　an　energy-intensive　manufacturing　system,　in　which　glass　is　produced　successively　in　cutting,　printing　and　tempering　stages.　The　problem　aims　to　simultaneously　optimize　makespan　and　the　total　electricity　cost　under　a　time-of-use　electricity　pricing　policy.　The　glass　production　has　to　respect　the　following　environments:　(i)　the　cutting　and　printing　operations　are　processed　in　parallel　machine　environments;　(ii)　the　tempering　operation　is　processed　on　a　batch　machine;　(iii)　machine　eligibility　and　setup　time　have　to　be　considered　in　the　cutting　and　printing　stages;　(iv)　the　whole　manufacturing　system　is　under　a　time-of-use　electricity　pricing　policy.　For　the　problem,　an　integer　programming　model　is　firstly　proposed　and　shown　to　be　strongly　NP-hard.　Then　a　model-based　heuristic　is　adopted　and　a　bi-objective　differential　evolution　algorithm　(BODE)　is　devised　based　on　problem　features.　Computational　experiments　on　randomly　generated　instances　demonstrated　that　the　BODE　outperforms　the　model-based　heuristic　in　terms　of　computation　time　and　solution　quality.　Moreover,　with　mild　increase　on　computation　burden,　the　BODE　significantly　outperforms　the　classic　NSGA　II　in　terms　of　solution　quality.　©　2018　Elsevier　Ltd