Scientists' research into fruit fly movement will affect fields such as medicine and driverless numbers.

Fruit fly larvae have an independent nervous system in the chest and abdomen.
this decentralized control mechanism can actually help them move very efficiently. the benefits of
central systems command and control of complex systems are often evident.
, for example, control our eyes and legs through the brain, so we can go straight.
such systems, whether natural or man-made, occasionally have serious vulnerabilities.
significant damage in important areas of its socio-economic activities during the recent suspension of some governments in the United States.
this vividly illustrates the consequences of a country's central control unit suddenly shutting down.
the death of brain cells in neurodegenerative diseases is increasingly destroying the body's controller (brain), showing that the biological system has the same weakness.
avoid the central system mechanism has its advantages.
lack of a "brain" to guide behavior means that the absence of individual components has little effect on collective behavior.
a team of researchers led by Neil Johnson, a physicist at George Washington University, developed a decentralized system model that successfully simulated the movement of fruit fly larvae.
the study was published February 6 in the journal Science Advances.
studies show that when a single part of a model is weak, it performs best - the simpler the component, the better the overall system.
by contrast, the components of a centralized system are better functional only if they are improved.
the researchers believe the findings could affect a wide range of areas, from the design of driverless cars to the treatment of neurological diseases to the treatment of neurological conditions to all aspects of the organization, and may even have implications for understanding evolutionary processes. Examples of
decentralized control are common in natural organisms and are found in bacteria, mucus bacteria and ant colonies.
Johnson was inspired by the observation: Because the neural circuits of fruit fly larvae are so simple, the various parts of the fruit fly larvae move in semi-independent ways during exercise.
this is an example of decentralized control in a single organism, as opposed to the "group intelligence" embodied in bees or other animals with collective behaviour.
although there is no central system to coordinate the various parts of the body of fruit fly larvae, they always achieve the goal of moving toward the desired temperature -- a process known as warming.
fruit fly larvae creep forward by shrinking the body.
when the larvae part of their body stretches and the other part shrinks, they turn.
thermal neurons determine the movement of various parts of the fruit fly's body, and the combined effects of these movements determine their steering angle. "The coordinated movement of the larvae is similar to the coordinated movement of the population to reach the exit,"
Johnson said.
is not that people call each other to tell each other to go to the exit, but according to outside information, this kind of collective gathering behavior will naturally occur.
" the researchers created a mathematical model that uses separate components to reproduce the movement of larvae, which store the results of past movements in memory (if the results show that the model is well aligned with the target direction, it is defined as 1, otherwise defined as 0).
each component selects the next action (turn left or right) based on past results by referring to a set of "policies" that link different sets of past results to different turning directions.
the researchers assigned different subsets of all possible strategies to different model components (corresponding to the semi-independent parts of the larvae) to some extent changing their behavior -- each model agent chose the best strategy at each step.
the team found that the creeping trajectories produced by the model looked very much like the real data of larvae crawling, leading them to believe they had captured some of the essence of the real system. David Wolpert, a mathematician at the Santa Fe Institute in
who was not involved in the study, said: "It's great to match the crawling trajectory of fruit flies.
shows that we have taken a good step forward in understanding these issues.
" this key finding is related to changes in the memory size of components.
the model behaves poorly when memory capacity is very small -- but its performance slows down when the capacity exceeds a certain size.
if the components of a system become too smart (in this case , the "m" is getting bigger and smarter), then the system will be further and further away from the intended target (d zone).
photo source: Pedro D. Manrique researchers used the "group/anti-group" theory to explain this result, which is a mathematical description of how individual components form a group with consistent behavior.
when the memory capacity is very small, a large number of components are formed, moving in the same direction.
they make a big turn, then suddenly turn in the other direction, producing exaggerated glyph movements.
if these components have too much memory, the group will be trapped by long-standing strategies determined by past results, without adequately considering the immediate message that they are off course.
there is a best size between these two extremes, the medium-sized groups that use the opposite strategy, just as half of the boats on both sides of the boat are paddling. "When you increase your memory, it's like overthinking," johnson said,
.
too much history reinforces the prejudices of the past.
", says Mr. Walbert, sometimes a single component has a similar effect when dealing with problems. "When people predict the stock market (based on the values of the past), they are careful not to pay much attention to the views of the past, " he says
.
it's more difficult to learn because too much attention to the past can bring clutter.
the team claims that the work could provide a new way of thinking about how evolution can move from natural, decentralized designs, such as bacteria, to living things like using centralized design.
this means that in a decentralized design, the "smart" degree of components can be limited without switching to a centralized design.
the team's next step is to study how the neural circuits that destroy fruit fly larvae with lasers , a bit like some oarsmen who lose function when paddling, affect movement.
team also wanted to explore the model's behavior by styming two oarsmen together, or inserting a super-memory rower in the middle of a stupid rower.
eventually, Johnson hopes to seek possible medical implications from the model.
future research will explore whether providing limited feedback to certain muscles in diseases such as Parkinson's can help suppress tremors caused by damage disassociated brain control signals. "We will consider applying for funding to complete this study accurately in the case of general motor neurone disease," johnson said,
.
we don't know if it's feasible, but in my opinion, we've at least proved it's theoretically feasible.
"Other areas that may be used in this study include driverless car design and organizational performance."
, however, Mr. Walbert was cautious.
he said the study did not compare the model with any other model, so there were few aspects of decentralized control that were better than centralized control.
noted that engineering systems can be reduced by simple replication.
but there are also situations where a group of robot soldiers, as a whole, need to be silent on the radio when they perform a special mission.
"Robots do not allow communication, so they have to run in a decentralized manner," he said, noting that the results suggest that, as engineers, at least the need to limit the cognitive abilities of robots in order to achieve the overall goal should be considered.
" Source: Prospect.com.