Re-tasking balloons in a balloon network based on expected failure modes of balloons

A computer-implemented method comprising:

• determining at a first time that a first balloon in a balloon network has a first predicted failure mode that corresponds to a first set of tasks, wherein the balloon network comprises at least a first sub-fleet of balloons and a second sub-fleet of balloons, wherein the first sub-fleet is assigned a first set of one or more tasks and the second sub-fleet is assigned a second set of one or more tasks;
• based on determining that the first balloon has the first predicted failure mode that corresponds to the first set of tasks, assigning the first balloon to the first sub-fleet of balloons;
• determining at a later second time that the first balloon has a second predicted failure mode that corresponds to the second set of tasks, wherein the second predicted failure mode comprises a different type of predicted balloon failure than the first predicted failure mode;
• determining that the second predicted failure mode is more likely to occur earlier than the first predicted failure mode;
• and
• in response to determining that the second predicted failure mode is more likely to occur earlier than the first predicted failure mode, reassigning the first balloon from the first sub-fleet of balloons to the second sub-fleet of balloons.

A system comprising:

• a plurality of sub-fleets of balloons, wherein each sub-fleet of balloons is assigned a corresponding set of one or more tasks within a balloon network;
• and
• a control system configured to:
• determine that a first balloon initially has a first predicted failure mode that corresponds to a first set of tasks;
• based on determining that the first balloon has a first predicted failure mode that corresponds to the first set of tasks, assign the first balloon to a first sub-fleet of balloons that is assigned the first set of tasks within the balloon network;
• periodically determine that the first balloon has a different predicted failure mode that corresponds to a set of tasks assigned to a different sub-fleet of balloons within the balloon network;
• determine that the different predicted failure mode is more likely to occur earlier than any other previously determined failure mode;
• and
• in response to determining that the different predicted failure mode is more likely to occur earlier than any other previously determined failure mode, reassign the first balloon to the different sub-fleet of balloons.

A non-transitory computer readable storage medium having stored therein instructions, that when executed by a computing device, cause the computing device to perform functions comprising:

• determining at a first time that a first balloon in a balloon network has a first predicted failure mode that corresponds to a first set of tasks, wherein the balloon network comprises at least a first sub-fleet of balloons and a second sub-fleet of balloons, wherein the first sub-fleet is assigned a first set of one or more tasks and the second sub-fleet is assigned a second set of one or more tasks;
• based on determining that the first balloon has the first predicted failure mode that corresponds to the first set of tasks, assigning the first balloon to the first sub-fleet of balloons;
• determining at a later second time that the first balloon has a second predicted failure mode that corresponds to the second set of tasks, wherein the second predicted failure mode comprises a different type of predicted balloon failure than the first predicted failure mode;
• determining that the second predicted failure mode is more likely to occur earlier than the first predicted failure mode;
• and
• in response to determining that the second predicted failure mode is more likely to occur earlier than the first predicted failure mode, reassigning the first balloon from the first sub-fleet of balloons to the second sub-fleet of balloons.