{
  "id": "brachistochrone-problem",
  "title": "Brachistochrone Problem",
  "lang": "en",
  "source_lang": "en",
  "summary": "The brachistochrone problem asks which curve allows a particle sliding under gravity to travel between two fixed points in the least time; its classical solution is a cycloid.",
  "category": "mathematics",
  "subcategory": "analysis",
  "type": "concept",
  "aliases": [
    "Problem of quickest descent",
    "Least-time descent problem"
  ],
  "generated_at": "2026-11-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/brachistochrone-problem/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/brachistochrone-problem/",
      "json": "https://aiwiki.page/zh/brachistochrone-problem.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The brachistochrone problem is a problem in the calculus of variations that asks for the curve along which a particle, released from rest and sliding without friction under uniform gravity, reaches a specified lower point in the shortest possible time. When the endpoints are not vertically aligned, the solution is an arc of an inverted cycloid, the curve traced by a point on the rim of a circle rolling along a straight line. The problem illustrates a fundamental distinction between shortest distance and shortest travel time: a longer path can be faster if it allows the particle to gain speed earlier."
    },
    {
      "heading": "Classical formulation",
      "text": "Consider two points in a vertical plane,\n\n\\[A=(0,0),\\qquad B=(X,Y),\\]\n\nwith \\(X>0\\) and \\(Y>0\\), where the vertical coordinate \\(y\\) is measured downward from the starting point. A particle begins at \\(A\\) with zero speed and moves along a fixed track to \\(B\\).\n\nThe classical assumptions are:\n\n- gravitational acceleration \\(g\\) is constant;\n- friction and air resistance are absent;\n- the particle slides rather than rolls;\n- the track constrains its motion without doing work;\n- there are no obstacles or restrictions on track depth.\n\nBy conservation of mechanical energy,\n\n\\[\\frac12 mv^2=mgy,\\]\n\nso its speed at depth \\(y\\) is\n\n\\[v=\\sqrt{2gy}.\\]\n\nThe particle’s mass \\(m\\) therefore has no effect on the optimal curve or the travel time.\n\nFor a track described as a graph \\(y=y(x)\\), the arc-length element is\n\n\\[ds=\\sqrt{1+(y')^2}\\,dx.\\]\n\nConsequently, the time of descent is the integral\n\n\\[T[y]=\\int_0^X \\frac{\\sqrt{1+(y')^2}}{\\sqrt{2gy}}\\,dx.\\]\n\nThe task is to minimize this expression over admissible curves satisfying the endpoint conditions. Unlike an elementary optimization problem, in which the unknowns are numbers, the unknown here is an entire curve.\n\nThe factor \\(1/\\sqrt{y}\\) is singular at the starting point because the particle initially has zero speed. This does not make the optimal travel time infinite: the time integral along the cycloid converges."
    },
    {
      "heading": "Why a straight line is not fastest",
      "text": "The straight segment joining the endpoints has the shortest length, but not the shortest descent time. Its relatively shallow initial slope may keep the particle moving slowly over a substantial part of the journey.\n\nA faster track begins more steeply, converting gravitational potential energy into kinetic energy earlier. It then turns toward the destination while the particle is moving rapidly. The optimal curve balances the additional distance against this increase in speed.\n\nThe curve need not descend continuously toward the endpoint. If the destination lies sufficiently far to the side relative to its depth, the fastest track first falls below the destination, then rises to meet it. During the final ascent the particle slows, but its earlier gain in speed more than compensates for the extra distance."
    },
    {
      "heading": "Variational derivation",
      "text": "Ignoring the constant factor \\(1/\\sqrt{2g}\\), define\n\n\\[F(y,y')=\\sqrt{\\frac{1+(y')^2}{y}}.\\]\n\nA sufficiently smooth minimizing curve must satisfy the Euler–Lagrange equation wherever \\(y>0\\).\n\nSince \\(F\\) has no explicit dependence on \\(x\\), the Euler–Lagrange equation admits the first integral known as the Beltrami identity:\n\n\\[F-y'\\frac{\\partial F}{\\partial y'}=C.\\]\n\nSubstitution gives\n\n\\[\\frac{1}{\\sqrt{y}\\sqrt{1+(y')^2}}=C.\\]\n\nWriting the positive constant in terms of a length \\(a\\) yields\n\n\\[y\\bigl(1+(y')^2\\bigr)=2a.\\]\n\nThis differential equation is solved parametrically by\n\n\\[x=a(\\theta-\\sin\\theta),\\qquad y=a(1-\\cos\\theta).\\]\n\nIndeed,\n\n\\[\\frac{dy}{dx} =\\frac{\\sin\\theta}{1-\\cos\\theta} =\\cot\\frac{\\theta}{2},\\]\n\nand hence\n\n\\[y\\bigl(1+(y')^2\\bigr) =a(1-\\cos\\theta)\\csc^2\\frac{\\theta}{2} =2a.\\]\n\nThese are the equations of a cycloid generated by a circle of radius \\(a\\), with its initial cusp at \\(A\\) and its arch extending downward.\n\nThe variational calculation identifies a stationary candidate; the Euler–Lagrange equation alone is not a general proof of global minimality. Establishing that the appropriate cycloidal segment is the least-time path also requires a sufficiency argument. The classical result is that the segment lying within the first arch solves the unrestricted frictionless problem."
    },
    {
      "heading": "Determining the cycloid from the endpoints",
      "text": "Let \\(\\theta_1\\) be the parameter value at the destination. The endpoint conditions are\n\n\\[X=a(\\theta_1-\\sin\\theta_1),\\qquad Y=a(1-\\cos\\theta_1).\\]\n\nEliminating \\(a\\) gives\n\n\\[\\frac{X}{Y} = \\frac{\\theta_1-\\sin\\theta_1}{1-\\cos\\theta_1}.\\]\n\nFor \\(X,Y>0\\), this equation determines a unique value\n\n\\[0<\\theta_1<2\\pi,\\]\n\nafter which\n\n\\[a=\\frac{Y}{1-\\cos\\theta_1}.\\]\n\nThe endpoint equation is transcendental and is generally solved using numerical methods.\n\nThe parameter \\(\\theta_1\\) also determines the track’s shape:\n\n- If \\(0<\\theta_1<\\pi\\), the particle descends throughout the journey.\n- If \\(\\theta_1=\\pi\\), the destination is the lowest point of the cycloid.\n- If \\(\\pi<\\theta_1<2\\pi\\), the particle passes below the destination and then ascends.\n\nThe transition occurs when\n\n\\[\\frac{X}{Y}=\\frac{\\pi}{2}.\\]\n\nAt the start, the cycloid has a vertical tangent. The first-arch solution has no interior cusp, even when it includes a final ascending portion. If the endpoints are vertically aligned, the limiting solution is a straight vertical drop."
    },
    {
      "heading": "Travel time along the solution",
      "text": "Along the cycloid, for \\(0<\\theta<2\\pi\\),\n\n\\[ds=2a\\sin\\frac{\\theta}{2}\\,d\\theta,\\]\n\nwhile energy conservation gives\n\n\\[v=2\\sqrt{ag}\\sin\\frac{\\theta}{2}.\\]\n\nTherefore,\n\n\\[dt=\\frac{ds}{v}=\\sqrt{\\frac{a}{g}}\\,d\\theta,\\]\n\nand the minimum travel time is\n\n\\[T_{\\min}=\\theta_1\\sqrt{\\frac{a}{g}}.\\]\n\nThus the cycloidal parameter increases uniformly with time:\n\n\\[\\theta=t\\sqrt{\\frac{g}{a}}.\\]\n\nThe particle’s speed itself is not constant; it increases during descent and decreases during any subsequent ascent.\n\nFor comparison, the straight track has length\n\n\\[L=\\sqrt{X^2+Y^2}\\]\n\nand constant tangential acceleration \\(gY/L\\). Its descent time is\n\n\\[T_{\\mathrm{straight}} =\\sqrt{\\frac{2(X^2+Y^2)}{gY}}.\\]\n\nFor a fixed endpoint geometry, both times scale as \\(g^{-1/2}\\). Changing the strength of a uniform gravitational field changes the duration but not the optimal track’s shape."
    },
    {
      "heading": "Optical interpretation",
      "text": "Johann Bernoulli’s solution used an analogy with light refraction. Under Fermat’s principle, a light ray follows a path of stationary travel time through a medium. Imagine a hypothetical medium in which light speed depends on depth according to\n\n\\[v(y)=\\sqrt{2gy}.\\]\n\nIts ray paths obey the same travel-time functional as the descending particle.\n\nDividing the medium into thin horizontal layers and applying Snell’s law shows that\n\n\\[\\frac{\\sin\\alpha}{v}\\]\n\nis constant, where \\(\\alpha\\) is the angle between the path and the vertical. Since\n\n\\[\\sin\\alpha=\\frac{1}{\\sqrt{1+(y')^2}},\\]\n\nthis produces the same first integral as the variational derivation.\n\nThe analogy also provides geometric intuition. Near the starting level, where the speed is small, the optimal path is nearly vertical. At greater depth, where the speed is higher, horizontal travel becomes less costly in time.\n\nMore abstractly, travel time defines a position-dependent length element,\n\n\\[d\\tau^2=\\frac{dx^2+dy^2}{2gy}.\\]\n\nAway from \\(y=0\\), the problem can therefore be viewed as finding a geodesic in a weighted geometry. The starting level is a singular boundary of that geometry."
    },
    {
      "heading": "Historical development",
      "text": "Galileo studied accelerated descent and compared straight and curved tracks in the seventeenth century, but did not obtain the correct general solution.\n\nJohann Bernoulli publicly posed the brachistochrone problem in Acta Eruditorum in June 1696. Solutions appeared in 1697, including those of Johann Bernoulli, Jakob Bernoulli, Gottfried Wilhelm Leibniz, and Isaac Newton.\n\nThe name derives from Greek words meaning “shortest” and “time.” The problem became a landmark in the development of the calculus of variations because it demanded the optimization of a curve rather than a finite collection of quantities. Later work by Leonhard Euler and Joseph-Louis Lagrange helped turn such problems into a systematic mathematical theory."
    },
    {
      "heading": "Relation to the tautochrone problem",
      "text": "The cycloid also solves the tautochrone problem, but the two questions are distinct:\n\n- The brachistochrone problem fixes two endpoints and asks which track gives the least travel time between them.\n- The tautochrone problem asks for a track on which a particle released from rest takes the same time to reach the bottom, regardless of its starting position.\n\nFor a frictionless particle released on a suitable cycloidal arch generated by a circle of radius \\(a\\), the time to reach the bottom is\n\n\\[\\pi\\sqrt{\\frac{a}{g}}.\\]\n\nChristiaan Huygens established the cycloid’s tautochrone property before Bernoulli posed the brachistochrone challenge.\n\nIn the brachistochrone problem, however, the destination need not be the bottom of the cycloid, and the travel time depends on its parameter value \\(\\theta_1\\)."
    },
    {
      "heading": "Extensions and limitations",
      "text": "The cycloidal solution depends on the idealized physical assumptions. Several changes produce different optimization problems:\n\n- Nonzero initial speed: If the initial speed is \\(v_0\\), then\n\n\\[v(y)=\\sqrt{v_0^2+2gy}.\\]\n\nThe effective zero-speed level lies above the starting point, so the track generally begins at a regular point of an appropriately shifted cycloid rather than at a cusp.\n- Rolling motion: For an ideal rigid body rolling without slipping, with fixed moment of inertia \\(I\\), mass \\(m\\), and rolling radius \\(r\\), rotational energy reduces its translational speed. In an idealized point-path model with constant \\(I/(mr^2)\\), this multiplies all travel times by the same factor and preserves the cycloidal optimum for the center-of-mass path. A finite body also introduces geometric and contact constraints.\n- Friction or drag: Dissipation generally makes speed depend on the previous motion, not merely on depth. The classical travel-time functional no longer applies unchanged.\n- Variable gravity: If gravitational acceleration varies with position, the relation between speed and position changes, and the optimum is generally not the classical cycloid.\n- Obstacles and track constraints: Limits on depth, slope, curvature, or acceleration alter the admissible curves. The vertical initial tangent of the ideal solution may be disallowed.\n\nThe brachistochrone remains a standard example in classical mechanics and variational analysis because it combines an exactly solvable optimization problem with a clear demonstration that distance, speed, and time must be optimized together."
    }
  ],
  "links": [
    "calculus-of-variations",
    "gravity",
    "energy",
    "integral",
    "mathematical-optimization",
    "euler-lagrange-equations",
    "differential-equation",
    "circle",
    "numerical-analysis",
    "fermats-principle",
    "snells-law",
    "geodesic",
    "gottfried-wilhelm-leibniz",
    "leonhard-euler",
    "classical-mechanics"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}