Page 335 -
P. 335

6.1 Basic Theory of Pseudodifferential Operators  319

                              To conclude this section, we now return to a subclass, the class of classi-
                           cal pseudodifferential operators, which is very important in connection with
                           elliptic boundary value problems and boundary integral equations.
                                                                               m
                           Definition 6.1.6. A pseudodifferential operator A ∈L (Ω) is said to
                           be classical if its complete symbol σ A has a representative in the class
                                                                           n
                                     n
                             m
                                                                  m
                           S S S (Ω × IR ),cf. (6.1.19). We denote by L (Ω × IR ) the set of all clas-
                             c
                                                                  c
                           sical pseudodifferential operators of order m.
                                                       n
                                              m
                           We remark that for L (Ω × IR ), the mapping
                                              c
                                        m
                                                                  n
                                                          m
                                                                                n
                                                                    :
                                       L (Ω) 
 A  → σ A ∈ S S S (Ω × IR ) S S S  −∞ (Ω × IR )
                                        c                 c
                           induces the isomorphism
                                     m   :  −∞        m        n  :  −∞      n
                                   L (Ω) L     (Ω) → S S S (Ω × IR ) S S S  (Ω × IR ) .  (6.1.39)
                                     c                c
                           Moreover, we have
                                                             (
                                                   −∞             m
                                                  L   (Ω)=      L (Ω) .                (6.1.40)
                                                                  c
                                                            m∈IR
                                               m
                           In this case, for A ∈L (Ω), the principal symbol σ mA has a representative
                                               c
                                                                     ξ

                                                 0
                                                             m
                                                a (x, ξ):= |ξ| a m x,                  (6.1.41)
                                                 m
                                                                     |ξ|
                                                    n
                           which belongs to C (Ω×(IR \{0})) and is positively homogeneous of degree
                                            ∞
                                                            0
                           m with respect to ξ. The function a (x, ξ) in (6.1.41) is called the homo-
                                                            m
                                                          m
                           geneous principal symbol of A ∈L (Ω). Note, in contrast to the principal
                                                          c
                           symbol of A defined in (6.1.31), the homogeneous principal symbol is only
                           a single function which represents the whole equivalence class in (6.1.31).
                           Correspondingly, if we denote by

                                                                        ξ
                                              0
                                                            m−j
                                             a m−j (x, ξ):= |ξ|  a m−j  x,             (6.1.42)
                                                                        |ξ|
                           the homogeneous parts of the asymptotic expansion of the classical symbol
                           σ mA , which have the properties
                               a 0  (x, ξ)  = a m−j (x, ξ)     for |ξ|  ≥ 1     and
                                m−j
                               a 0  (x, tξ)  = t m−j 0  (x, ξ)  for all t  > 0  and ξ  =0 ,
                                                  a
                                m−j                m−j
                                                                                  ∞
                                                                                     0
                           then σ mA may be represented asymptotically by the formal sum  a m−j (x, ξ).
                                                                                 j=0
   330   331   332   333   334   335   336   337   338   339   340