bogus backfilling in easy backfilling implementation

Robin BOEZENNEC reported on the mattermost (see https://framateam.org/batsim/pl/7rpyabwawffz3q8u3w9zwcm7ah and following messages) channel an incoherent schedule produced by the EASY backfill algorithm shipped in pybatsim-core.

The provided workload is the following:

{
  "description" :  "", 
  "command": "", 
  "date": "", 
  "jobs": [ 
    { "id" : "0_0_1", "profile" : "profile_0", "res" : 4, "subtime" : 0, "load_id" : 0, "essai_numero" : 0, "nb_times" : 1, "walltime" : 86690}, 
    { "id" : "1_0_1", "profile" : "profile_1", "res" : 32, "subtime" : 1, "load_id" : 1, "essai_numero" : 0, "nb_times" : 1, "walltime" : 65391}, 
    { "id" : "2_0_1", "profile" : "profile_2", "res" : 32, "subtime" : 2, "load_id" : 2, "essai_numero" : 0, "nb_times" : 1, "walltime" : 53070}, 
    { "id" : "3_0_1", "profile" : "profile_3", "res" : 128, "subtime" : 3, "load_id" : 3, "essai_numero" : 0, "nb_times" : 1, "walltime" : 4182}, 
    { "id" : "4_0_1", "profile" : "profile_4", "res" : 32, "subtime" : 4, "load_id" : 4, "essai_numero" : 0, "nb_times" : 1, "walltime" : 61432}, 
    { "id" : "5_0_1", "profile" : "profile_5", "res" : 32, "subtime" : 5, "load_id" : 5, "essai_numero" : 0, "nb_times" : 1, "walltime" : 59189}, 
    { "id" : "6_0_1", "profile" : "profile_6", "res" : 4, "subtime" : 6, "load_id" : 6, "essai_numero" : 0, "nb_times" : 1, "walltime" : 86522}, 
    { "id" : "7_0_1", "profile" : "profile_7", "res" : 128, "subtime" : 7, "load_id" : 7, "essai_numero" : 0, "nb_times" : 1, "walltime" : 2639}, 
    { "id" : "8_0_1", "profile" : "profile_8", "res" : 32, "subtime" : 8, "load_id" : 8, "essai_numero" : 0, "nb_times" : 1, "walltime" : 65260}, 
    { "id" : "9_0_1", "profile" : "profile_9", "res" : 128, "subtime" : 9, "load_id" : 9, "essai_numero" : 0, "nb_times" : 1, "walltime" : 1586}, 
    { "id" : "10_0_1", "profile" : "profile_10", "res" : 4, "subtime" : 10, "load_id" : 10, "essai_numero" : 0, "nb_times" : 1, "walltime" : 86650}, 
    { "id" : "11_0_1", "profile" : "profile_11", "res" : 32, "subtime" : 11, "load_id" : 11, "essai_numero" : 0, "nb_times" : 1, "walltime" : 59758}, 
    { "id" : "12_0_1", "profile" : "profile_12", "res" : 32, "subtime" : 12, "load_id" : 12, "essai_numero" : 0, "nb_times" : 1, "walltime" : 59056}, 
    { "id" : "13_0_1", "profile" : "profile_13", "res" : 32, "subtime" : 13, "load_id" : 13, "essai_numero" : 0, "nb_times" : 1, "walltime" : 62136}, 
    { "id" : "14_0_1", "profile" : "profile_14", "res" : 4, "subtime" : 14, "load_id" : 14, "essai_numero" : 0, "nb_times" : 1, "walltime" : 87042}, 
    { "id" : "15_0_1", "profile" : "profile_15", "res" : 128, "subtime" : 15, "load_id" : 15, "essai_numero" : 0, "nb_times" : 1, "walltime" : 4577}, 
    { "id" : "16_0_1", "profile" : "profile_16", "res" : 128, "subtime" : 16, "load_id" : 16, "essai_numero" : 0, "nb_times" : 1, "walltime" : 3434}
  ],
  "nb_res": 49152, 
  "profiles":
  {
    "profile_0" : {"type": "delay", "runtime" : 78809, "delay" : 78809, "np" : 4, "command" : "????"}, 
    "profile_1" : {"type": "delay", "runtime" : 59446, "delay" : 59446, "np" : 32, "command" : "????"}, 
    "profile_2" : {"type": "delay", "runtime" : 48245, "delay" : 48245, "np" : 32, "command" : "????"}, 
    "profile_3" : {"type": "delay", "runtime" : 3801, "delay" : 3801, "np" : 128, "command" : "????"}, 
    "profile_4" : {"type": "delay", "runtime" : 55847, "delay" : 55847, "np" : 32, "command" : "????"}, 
    "profile_5" : {"type": "delay", "runtime" : 53808, "delay" : 53808, "np" : 32, "command" : "????"}, 
    "profile_6" : {"type": "delay", "runtime" : 78656, "delay" : 78656, "np" : 4, "command" : "????"}, 
    "profile_7" : {"type": "delay", "runtime" : 2399, "delay" : 2399, "np" : 128, "command" : "????"}, 
    "profile_8" : {"type": "delay", "runtime" : 59327, "delay" : 59327, "np" : 32, "command" : "????"}, 
    "profile_9" : {"type": "delay", "runtime" : 1441, "delay" : 1441, "np" : 128, "command" : "????"}, 
    "profile_10" : {"type": "delay", "runtime" : 78772, "delay" : 78772, "np" : 4, "command" : "????"}, 
    "profile_11" : {"type": "delay", "runtime" : 54325, "delay" : 54325, "np" : 32, "command" : "????"}, 
    "profile_12" : {"type": "delay", "runtime" : 53687, "delay" : 53687, "np" : 32, "command" : "????"}, 
    "profile_13" : {"type": "delay", "runtime" : 56487, "delay" : 56487, "np" : 32, "command" : "????"}, 
    "profile_14" : {"type": "delay", "runtime" : 79129, "delay" : 79129, "np" : 4, "command" : "????"}, 
    "profile_15" : {"type": "delay", "runtime" : 4160, "delay" : 4160, "np" : 128, "command" : "????"}, 
    "profile_16" : {"type": "delay", "runtime" : 3121, "delay" : 3121, "np" : 128, "command" : "????"}
  }
}

and the platform used to run the experiment is as follows:

<?xml version='1.0'?>
<!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
<platform version="4.1">

<zone id="AS0" routing="Full">
    <cluster id="my_cluster_1" prefix="a" suffix="" radical="0-127"
        speed="1" bw="1" lat="1" bb_bw="1" core="1"
        bb_lat="1" />

    <cluster id="my_cluster_2" prefix="master_host" suffix="" radical="0-0"
        speed="1Gf" bw="125MBps" lat="50us" bb_bw="2.25GBps"
        bb_lat="500us">
        <prop id="role" value="master" />
    </cluster>

    <link id="backbone" bandwidth="1.25GBps" latency="500us" />

    <zoneRoute src="my_cluster_1" dst="my_cluster_2" gw_src="amy_cluster_1_router"
        gw_dst="master_hostmy_cluster_2_router">
        <link_ctn id="backbone" />
    </zoneRoute>
</zone>
</platform>

The schedule obtained when running pybatsim is bogus:

As a reference, here is the schedule produced by the batsched implementation: