C++ unqualified-id，其他帖子没有帮助

Question

我正在使用 linux 机器并使用 g++ 编译 .cpp 文件。我不断收到以下错误：

hillClimbing.cpp:6: error: expected unqualified-id before â[â token
hillClimbing.cpp:7: error: expected unqualified-id before â[â token
hillClimbing.cpp:8: error: expected unqualified-id before â[â token
hillClimbing.cpp:17: error: expected unqualified-id before â[â token

我一直在查看此站点上的类似帖子，但它们似乎与我的问题不符，我无法解决它。这是 .cpp 文件：

#include <utility>
#include "evaluateParams.h"

using namespace std;

    double[] optimizeParams(EvaluateParams eval, pair<double, double> ranges[], double ss[], double a, double e, bool hillClimb, bool ascent);
    double[] initParams(pair<double, double> ranges[]);
    double[] climbHill(double pos[], double stepSize[], double epsilon, bool findMax);

    evaluateParams evalMethod;
    double accel;

    //eval/evalMethod: defines method used to compare a set of params
    //ranges: list of ranges that each parameter can take
    //ppp: points per parameter, the number of points initialized is ppa^ranges.size()
    //hillClimb: indicates weather to enact the hillClimb algorithm or simply randomize
    double[] optimizeParams(evaluateParams eval, pair<double, double> ranges[], double ss[], double a, double e, bool hillClimb, bool ascent){
            evalMethod = eval;
            accel = a;

            double paramValues[] = initParams(ranges);
            if(hillClimb){
                    paramValues = climbHill(paramValues[i], ss, e, ascent);
            }
            return paramValues;
    }

    double[] initParams(pair<double, double> ranges[]){
            double values[ranges.size()];
            for(int j = 0; j < ranges.size(); j++){
                    double min = ranges[i].first;
                    double max = ranges[i].second;
                    double r = ((double) rand() / (RAND_MAX));
                    r *= max-min;
                    r += min;
                    values[i][j]=r;
            }
            return values;
    }

    double[] climbHill(double pos[], double stepSize[], double epsilon, bool findMax){
            double candidate[] = [-accel, -1/accel, 0, 1/accel, accel];
            while(true){ //may need to switch to a timer in case points get stuck
                    double init = eval.evaluate(pos);
                    for(int i = 0; i < pos.size(); i++){
                            int best = -1;
                            double bestEval = -1 * numeric_limits<double>::max();
                            for(int j = 0; j < candidate.size(); j++){
                                    pos[i] = pos[i] + stepSize[i]*candidate[j];
                                    double temp = eval.evaluate(tempPos);
                                    pos[i] = pos[i] - stepSize[i]*candidate[j];
                                    if(temp > bestScore){
                                            bestEval = temp;
                                            best = j;
                                    }
                            }
                            if(findMax){
                                    pos[i] = pos[i] + stepSize[i] * candidate[best];
                            }
                            else{
                                    pos[i] = pos[i] - stepSize[i] * candidate[best];
                            }
                            if(candidate[best] != 0){
                                    stepSize[i] = stepSize[i] * candidate[best];
                            }
                    }
                    if(eval.evaluate(pos)-init < epsilon){
                            return pos;
                    }
            }
    }

这里是 .h 文件：

class evaluateParams{
    public:
            evaluateParams(){}
            virtual double evaluate(double[]);

};

我看到的大多数其他帖子都是通过添加“;”来修复的在头文件的末尾，我已经完成了。

Answer 1

C++ 中的数组的声明方式与 Java 中的不同。此外，C++ 内置数组不是动态调整大小的。您可能应该用std::vector<double> 替换double[] 的所有用法：这是指定动态大小的对象数组的C++ 方式。

Answer 2

函数的返回类型可能没有定义为数组>所以这些声明是无效的

double[] optimizeParams(EvaluateParams eval, pair<double, double> ranges[], double ss[], double a, double e, bool hillClimb, bool ascent);
double[] initParams(pair<double, double> ranges[]);
double[] climbHill(double pos[], double stepSize[], double epsilon, bool findMax);

它们必须被声明为

double * optimizeParams(EvaluateParams eval, pair<double, double> ranges[], double ss[], double a, double e, bool hillClimb, bool ascent);
double * initParams(pair<double, double> ranges[]);
double * climbHill(double pos[], double stepSize[], double epsilon, bool findMax);

当您返回指向本地数组的第一个元素的指针时，程序的行为是未定义的。

此声明也无效

double paramValues[] = initParams(ranges);

数组可能不会以这种方式初始化。