Construction, the bound algebra, and the interval kernels of the elementary functions.
Attributes
Members list
Value members
Concrete methods
a + b, when at least one is an interval; None otherwise or when not representable.
a + b, when at least one is an interval; None otherwise or when not representable.
Value parameters
- a
-
the left operand
- b
-
the right operand
- env
-
the environment, for the exact tier's gcd policy
Attributes
- Returns
-
the sum's interval, or
None
Whether every point of a lies below every point of b (strict: <, else ≤), with the = tolerance: Some(true) when it holds for every pair, Some(false) when it fails for every pair, None (unknown) otherwise. At least one operand must be an interval.
Whether every point of a lies below every point of b (strict: <, else ≤), with the = tolerance: Some(true) when it holds for every pair, Some(false) when it fails for every pair, None (unknown) otherwise. At least one operand must be an interval.
Value parameters
- a
-
the left operand
- b
-
the right operand
- strict
-
<rather than≤ - tol
-
the tolerance within which two points count as equal
Attributes
- Returns
-
the verdict, or
Nonewhen it depends on the points chosen
Whether a relation between a and b is a question about points of an interval: one of them is an interval and both have real bounds (F_0101). A complex number, a matrix or a truth value has none, so a relation with one is not decided here at all — it stays symbolic, as it would without the interval, rather than being answered unknown. below and equal answer None for both cases, so a caller asks this first.
Whether a relation between a and b is a question about points of an interval: one of them is an interval and both have real bounds (F_0101). A complex number, a matrix or a truth value has none, so a relation with one is not decided here at all — it stays symbolic, as it would without the interval, rather than being answered unknown. below and equal answer None for both cases, so a caller asks this first.
Value parameters
- a
-
the left operand
- b
-
the right operand
Attributes
a / b, when at least one is an interval; None when the divisor contains zero, since the quotient is then two disjoint rays, not an interval.
a / b, when at least one is an interval; None when the divisor contains zero, since the quotient is then two disjoint rays, not an interval.
Value parameters
- a
-
the dividend
- b
-
the divisor
- env
-
the environment, for the exact tier's gcd policy
Attributes
- Returns
-
the quotient's interval, or
None
Whether a = b for every pair of points (Some(true)), for none (Some(false): the intervals are apart by more than tol), or neither (None, unknown).
Whether a = b for every pair of points (Some(true)), for none (Some(false): the intervals are apart by more than tol), or neither (None, unknown).
Value parameters
- a
-
the left operand
- b
-
the right operand
- tol
-
the
=tolerance,0.5·10⁻ᵖ
Attributes
- Returns
-
the verdict, or
Nonewhen unknown or not comparable
The interval image of a named elementary function, or None when the function has no interval rule here or the interval leaves its domain (a pole of tan, ln reaching zero, asin beyond ±1).
The interval image of a named elementary function, or None when the function has no interval rule here or the interval leaves its domain (a pole of tan, ln reaching zero, asin beyond ±1).
Value parameters
- a
-
the argument, an interval
- env
-
supplies the working precision for an exact interval
- name
-
the function's grammar name:
exp ln sin cos tan asin acos atan sinh cosh tanh abs
Attributes
- Returns
-
the image, or
None
log_base(a) as ln(a) / ln(base), either of them an interval: a number's logarithm is itself enclosed, so the quotient stays rigorous. None outside the domain, and for a base whose logarithm interval contains zero (a base at or around 1).
log_base(a) as ln(a) / ln(base), either of them an interval: a number's logarithm is itself enclosed, so the quotient stays rigorous. None outside the domain, and for a base whose logarithm interval contains zero (a base at or around 1).
Value parameters
- a
-
the argument
- base
-
the base
- env
-
supplies the working precision for exact operands
Attributes
- Returns
-
the enclosure, or
None
a · b, when at least one is an interval.
a · b, when at least one is an interval.
Value parameters
- a
-
the left operand
- b
-
the right operand
- env
-
the environment, for the exact tier's gcd policy
Attributes
- Returns
-
the product's interval, or
Nonewhen neither is an interval or it is not representable
An interval from two endpoints, or None when they do not make one: an endpoint that is not a real number, a non-finite or NaN one, or lo > hi. Equal endpoints give the number itself.
An interval from two endpoints, or None when they do not make one: an endpoint that is not a real number, a non-finite or NaN one, or lo > hi. Equal endpoints give the number itself.
Value parameters
- hi
-
the upper endpoint
- lo
-
the lower endpoint
Attributes
- Returns
-
the interval, the number, or
None
a ^ e for an interval a and a numeric exponent; None for an interval exponent, a negative power of an interval containing zero, a fractional power of an interval reaching below zero, or a result that is not representable.
a ^ e for an interval a and a numeric exponent; None for an interval exponent, a negative power of an interval containing zero, a fractional power of an interval reaching below zero, or a result that is not representable.
An integer power keeps the even/odd distinction ([−2, 1]² = [0, 4], not [1, 4]); a fractional one is monotone on [0, ∞).
Value parameters
- a
-
the base, an interval
- e
-
the exponent, a number
- env
-
the environment, for the exact tier's gcd policy and working precision
Attributes
- Returns
-
the power's interval, or
None