#include<bits/stdc++.h>
#pragma GCC optimize("Ofast")
//#pragma GCC target("avx2")
#pragma GCC optimize("unroll-loops")
using namespace std;
//#include<boost/multiprecision/cpp_int.hpp>
//#include<boost/multiprecision/cpp_dec_float.hpp>
//namespace mp=boost::multiprecision;
//#define mulint mp::cpp_int
//#define mulfloat mp::cpp_dec_float_100
struct __INIT{__INIT(){cin.tie(0);ios::sync_with_stdio(false);cout<<fixed<<setprecision(15);}} __init;
//#define INF (1<<30)
#define LINF (lint)(1LL<<56)
#define MINF (lint)(2e18)
#define endl "\n"
#define rep(i,n) for(lint (i)=0;(i)<(n);(i)++)
#define reprev(i,n) for(lint (i)=(n-1);(i)>=0;(i)--)
#define flc(x) __builtin_popcountll(x)
#define pint pair<int,int>
#define pdouble pair<double,double>
#define plint pair<lint,lint>
#define fi first
#define se second
#define all(x) x.begin(),x.end()
//#define vec vector<lint>
#define nep(x) next_permutation(all(x))
typedef long long lint;
int dx[8]={1,1,0,-1,-1,-1,0,1};
int dy[8]={0,1,1,1,0,-1,-1,-1};
const int MAX_N=3e5+5;
template<class T>bool chmax(T &a,const T &b){if(a<b){a=b;return 1;}return 0;}
template<class T>bool chmin(T &a,const T &b){if(b<a){a=b;return 1;}return 0;}
//vector<int> bucket[MAX_N/1000];
//constexpr int MOD=1000000007;
constexpr int MOD=998244353;
//typedef __int128_t llint;
#include<atcoder/all>
using namespace atcoder;

int main(void){
    lint N;
    cin >> N;
    lint P[N];
    rep(i,N) cin >> P[i];
    lint ans=0;
    lint pow2[N+1];
    pow2[0]=1;
    rep(i,N){
        pow2[i+1]=pow2[i]*2;
        pow2[i+1]%=MOD;
    }
    rep(i,N){
        for(int j=i+1;j<N;j++) ans+=(P[i]>P[j])*(pow2[N-1]-pow2[N-1-(j-i)]+MOD);
        ans%=MOD;
    }
    cout << ans << endl;
}